Linking Mineralogy to Lithogeochemistry in the Highland Valley Copper District: Implications for Porphyry Copper Footprints

The Highland Valley Copper porphyry deposits, hosted in the Late Triassic Guichon Creek batholith in the Canadian Cordillera, are unusual in that some of them formed at depths of at least 4 to 5 km in cogenetic host rocks. Enrichments in ore and pathfinder elements are generally limited to a few hundred meters beyond the pit areas, and the peripheral alteration is restricted to narrow (1–3 cm) halos around a low density of prehnite and/or epidote veinlets. It is, therefore, challenging to recognize the alteration footprint peripheral to the porphyry Cu systems. Here, we document a workflow to maximize the use of lithogeochemical data in measuring changes in mineralogy and material transfer related to porphyry formation by linking whole-rock analyses to observed alteration mineralogy at the hand specimen and deposit scale. Alteration facies and domains were determined from mapping, feldspar staining, and shortwave infrared imaging and include (1) K-feldspar halos (potassic alteration), (2) epidote veins with K-feldspar–destructive albite halos (sodic-calcic alteration), (3) quartz and coarse-grained muscovite veins and halos and fine-grained white-mica–chlorite veins and halos (white-mica–chlorite alteration), and two subfacies of propylitic alteration comprising (4) prehnite veinlets with white-mica–chlorite-prehnite halos, and (5) veins of epidote ± prehnite with halos of chlorite and patchy K-feldspar. Well-developed, feldspar-destructive, white-mica alteration is indicated by (2[Ca-C] + N + K)/Al values <0.85, depletion in CaO and Na2O, enrichment in K2O, and localized SiO2 addition and is spatially limited to within ~200 m of porphyry Cu mineralization. Localized K2O, Fe2O3, and depletion in Cu, and some enrichment in Na2O and CaO, occurs in sodic-calcic domains that form a large (~34 km2) nonconcentric footprint outboard of well-mineralized and proximal zones enriched in K. Water and magmatic CO2-rich propylitic and sodic-calcic–altered rocks form the largest lithogeochemical footprint to the mineralization in the Highland Valley Copper district (~60 km2). Calcite in the footprint is interpreted to have formed via phase separation of CO2 from a late-stage magmatic volatile phase. Several observations from this study are transferable to other porphyry systems and have implications for porphyry Cu exploration. Feldspar staining and shortwave infrared imaging highlight weak and cryptic alteration that did not cause sufficient material transfer to be confidently distinguished from protolith lithogeochemical compositions. Prehnite can be a key mineral phase in propylitic alteration related to porphyry genesis, and its presence can be predicted based on host-rock composition. Sodic-calcic alteration depletes the protolith in Fe (and magnetite) and, therefore, will impact petrophysical and geophysical characteristics of the system. Wholerock loss on ignition and C and S analyses can be used to map enrichment in water and CO2 in altered rocks, and together these form a large porphyry footprint that extends beyond domains of enrichment in ore and pathfinder elements and of pronounced alkali metasomatism.


Introduction
Calc-alkalic porphyry deposits typically have a predictable spatial and temporal distribution of alteration minerals that is the result of the thermal, mechanical, and chemical evolution of the magmatic-hydrothermal system (Seedorff et al., 2005;Sillitoe, 2010). The alteration zones are diagnostic of the fluid composition and temperatures in which they formed and are commonly depicted as volumes of altered rocks in and around Cu mineralization. The deep and central part of the hydrothermal system is characterized by potassic alteration that zones upward to sericitic and then advanced argillic alteration (Seedorff et al., 2005;Sillitoe, 2010). Additionally, at deep levels, potassic alteration can zone laterally into sodic-calcic or propylitic alteration (Seedorff et al., 2008;Halley et al., 2015), whereas at intermediate and shallow levels the peripheral alteration comprises propylitic and intermediate argillic assemblages (Seedorff et al., 2005;Sillitoe, 2010). Thus, the plan view distribution of the alteration facies (i.e., the porphyry footprint, which is commonly mapped at surface) is partly a function of erosional level. A challenge for explorationists is that the peripheral parts of porphyry Cu systems are most commonly characterized by green rock propylitic alteration that lacks the diagnostic enrichment in pathfinder elements that occurs with sulfide mineralization and/or sericite alteration (Halley et al., 2015). Propylitic alteration manifests as low-to moderate-intensity replacement of primary feldspars and mafic minerals by epidote-calcitechlorite (± albite, actinolite, pyrite, hematite, prehnite, and zeolites) and is typically associated with veins of epidote ± chlorite ± carbonate (Meyer and Hemley, 1967;Cooke et al., 2014a, b). Propylitic alteration can extend for several kilometers away from the well-mineralized potassically and hydrolytically altered zones (Seedorff et al., 2005;Cooke et al., 2014a) and, thus, forms the largest footprint. Sodic-calcic or sodic alteration has also been recognized in the peripheral and deep (root) parts of several Laramide porphyry systems in the western part of the United States Seedorff et al., 2008;Runyon et al., 2019) and in some porphyry Cu systems in British Colombia (Jago et al., 2014;Byrne et al., 2020). Sodic-calcic alteration is characterized by the replacement of K-feldspar and/or plagioclase with albite ± fine-grained white mica and the alteration of mafic minerals to actinolite ± epidote, chlorite, and titanite, which generally occurs with veins of epidote or actinolite Runyon et al., 2019;Byrne et al., 2020).
Few studies have tested for lithogeochemical changes in the altered rocks peripheral to Cu centers and pyrite halos, particularly at the district scale (e.g., Urqueta et al., 2009;Greenlaw, 2014;Ahmed et al., 2019); nevertheless, some commonalities emerge. By many definitions, propylitic alteration involves material transfer of H2O, CO2, and localized S to the altered rocks but limited hydrolysis or alkali exchange metasomatism (Meyer and Hemley, 1967;Ulrich and Heinrich, 2001;Urqueta et al., 2009). In contrast to propylitic facies, sodic-calcic and sodic alteration causes exchange of K for Na, with varying degrees of Ca and CO2 addition, and can be accompanied by formation of hydrous minerals (Carten, 1986;. There is a paucity of studies that assess the district-scale spatial variation of the volatile components (H2O, OH, CO2) in the porphyry footprint (e.g., Djouka-Fonkwe et al., 2012).
Four major porphyry Cu-(Mo) deposits occur in the Highland Valley Copper district in British Columbia and constitute Canada's largest and longest-operating Cu mine ( Fig. 1; Ydon, 2007;Byrne et al., 2013;Sillitoe, 2013). The Highland Valley Copper porphyry systems are hosted in the Late Triassic Guichon Creek batholith and are unusual in that some of them formed at depths of at least 4 to 5 km (D'Angelo, 2016;D'Angelo et al., 2017). Furthermore, ore (Cu, Mo, Ag) and pathfinder element (e.g., As, Bi, W, Te) concentrations drop to background levels just a few hundred meters beyond the porphyry Cu mineralization (and pit margins), and alteration in the plutonic host rock is restricted to generally narrow 1-to 3-cm halos around a low density of prehnite and/or epidote veinlets (Olade andFletcher, 1975, 1976;Lesage et al., 2016). This is a particularly challenging style of footprint to explore, because outside of the mine pit areas the geophysical re-sponse of the porphyry systems is subdued (Ager et al., 1972;Roy and Clowes, 2000;Byrne et al., 2019;Vallée et al., 2019). The question, then, is how to use lithogeochemistry, in conjunction with mapping, to the best effect in such a challenging environment. As part of the Natural Sciences and Engineering Research Council (NSERC)-Canadian Mining Innovation Council Mineral Exploration Footprints Research Network (Lesher et al., 2017) the footprint of the Highland Valley Copper porphyry district has been investigated. We used feldspar staining, short-wave infrared (SWIR) spectral imaging, and optical microscopy to characterize vein and alteration mineralogy in rock slabs and to categorize the corresponding lithogeochemical samples. We applied molar element ratio (MER) diagrams (Stanley and Madeisky, 1994) and calculate element gains and losses (Grant, 1986). This allowed us to quantify mineralogical changes and material transfer and, thus, evaluate alteration processes.
By linking the observed alteration mineralogy to the lithogeochemical data, we tested how sensitive whole-rock determinations are in detecting changes in mineralogy (and material transfer) in the porphyry footprint. We show that the host-rock composition influences the alteration mineralogy, that material transfer occurred in the green rock outboard of mineralization, and that feldspar staining can identify subtle alteration that cannot be determined from lithogeochemistry. Additionally, we examine the paragenetic origin of secondary calcite and how it contributes to the porphyry footprint. Therefore, this contribution outlines methodologies and approaches that will be useful in elucidating the hydrothermal footprint in porphyry Cu systems globally.

Regional geology
The Quesnel terrane in the Canadian Cordillera is characterized by Mesozoic island-arc assemblages, including the Triassic Nicola Group (Coney et al., 1980). The Nicola Group (Fig. 1B) comprises mafic to intermediate volcaniclastic and volcanogenic sedimentary rocks, sandstone, limestone, layered siltstone, calcareous siltstone, and several chert horizons that were predominantly deposited in a submarine setting (McMillan, 1976(McMillan, , 1977McMillan et al., 2009). The Triassic Guichon Creek batholith, the host of the Highland Valley Copper deposits, is an I-type composite batholith (D'Angelo et al., 2017) that intruded Nicola Group rocks along the western margin of the Quesnel terrane (Logan and Mihalynuk, 2014).

District-scale geologic framework
The Guichon Creek batholith formed from multiple pulses of magmatism, is compositionally and concentrically zoned, with an oval shape measuring approximately 60 × 25 km, and is elongate to the northwest (Fig. 1). The initial magmatic pulse consists of mafic mineral-rich quartz-diorite, monzogabbro, and granodiorites (the Border, Guichon, and Chataway units), which occur at the margin of the batholith. The second pulse comprises mafic, mineral-poor, quartz-rich granodiorites and monzogranite facies (Bethlehem, Skeena, and Bethsaida;Fig. 1B;D'Angelo et al., 2017;Lee et al., 2020). Syn-to latemineralization stocks and dikes occur at some of the porphyry centers (Byrne et al., 2013. The Highland Valley Cop- per district is situated within the younger, more felsic facies in the core of the batholith and comprises the Bethlehem, J.A., Valley-Lornex, and Highmont porphyry Cu-(Mo) systems ( Fig. 2A). Postmineralization faulting along the Lornex fault offset the once contiguous Valley-Lornex system (McMillan, 1974;Hollister et al., 1975). At least two main stages of porphyry Cu mineralization are evident in the Highland Valley Copper district. The first was at the Bethlehem center at ~209 Ma and is associated with magmatic-hydrothermal breccias and cogenetic porphyry dikes that were emplaced in a dome-shaped feature at the Guichon-Bethlehem granodiorite contact (Briskey and Bellamy, 1976;Briskey, 1980;Byrne et al., 2013;D'Angelo et al., 2017). Bornite and chalcopyrite mineralization is focused in the breccia and synmineralization dikes and occurs with potassic alteration that zones up, and outward, to fine-grained white-mica-chlorite veins and alteration. Mineralization also occurs in D-type veins peripheral to the Cu center (Table 1). Late-to postmineralization stocks, compositionally similar to the Skeena facies, crosscut mineralized breccia at depth in the Bethlehem porphyry center (Byrne et al., 2013). Sodiccalcic alteration overprints potassic facies and Cu mineralization at Bethlehem and is spatially associated with postmineral Skeena stocks emplaced beneath the porphyry center . The timing of alteration and mineralization at J.A. is undefined; however, the mineralization and alteration style suggests it is akin to Bethlehem (McMillan, 1985b). Mineralization at Bethlehem was followed by the intrusion and crystallization of the Skeena and Bethsaida facies and, subsequently, the formation of the Valley-Lornex and Highmont porphyry centers between 208 and 207 Ma (D'Angelo et al., 2017). The Valley-Lornex porphyry is hosted in Skeena and Bethsaida granodiorite. Weak chalcopyrite mineralization occurs with early-stage quartz veinlets with narrow K-feldspar halos and is centered on a core domain of Cu-barren quartz veins (Table 1). Main-stage Cu mineralization is represented by quartz-bornite-chalcopyrite ± molybdenite veins with halos of coarse-grained, gray-colored, Na-bearing muscovite (Table 1; Alva-Jimenez et al., 2020) and predominately overprints sodic-calcic alteration . Chalcopyrite mineralization also occurs in veins with fine-grained whitemica-chlorite halos at Valley-Lornex and Highmont (Byrne et al., 2013) and in structurally controlled D-type veins that are present up to ~1 km away from the porphyry centers (Table  1). Weak late-stage intermediate-argillic alteration occurs throughout the Valley system (Jambor and Delabio, 1978), but more intense clay alteration is found in fault zones.

Mapping and sample workflow
Outcrop mapping, core logging, and lithogeochemical sampling was conducted between 2013 and 2016. Mapped alteration captures domains of specific assemblages with similar vein densities or alteration intensity ( Fig. 2A).
Representative, coherent blocks with samples of fractures, veins (800-1,000 cm 3 ), and their alteration assemblages were taken at every available outcrop, typically at a spacing of 0.5 to 1.5 km apart (Fig. 2B). The sample blocks were portioned and slabbed using a rock saw to ensure weathered and organic material was removed before analyses. This methodology was intended to standardize the samples. A total of 867 samples underwent lithogeochemical analysis, feldspar staining, and SWIR spectral imaging. Lithogeochemical data for an additional 181 samples were provided by Teck Resources Limited and are utilized in the maps in this contribution.
Observations made from rock slabs, feldspar staining, and SWIR spectral imaging were integrated to define the alteration facies present in each sample (Table 1). Results were tabulated and linked to the corresponding lithogeochemical analysis (App. 1).

Lithogeochemical analysis
The lithogeochemical analyses were carried out at the Bureau Veritas Laboratories, Vancouver, Canada. A crushed sample split was fused with a lithium metaborate/tetraborate mix. The fused bead was digested using a four-acid solution, and the major element oxides were analyzed by inductively coupled plasma-emission spectroscopy (ICP-ES). Trace elements were measured on an inductively coupled plasma-mass spectrometer (ICP-MS). Loss on ignition (LOI) is reported as percent weight loss on a 1 g sample ignited at 1,000°C. Carbon and S were measured by LECO. Sample preparation, analytical digestion methods, and quality assurance and control procedures are provided in Appendix 2. Total C (wt %) was converted to CO2, and then the amount of CO2 and S in each analysis was subtracted from the LOI measurement to estimate the total mineral-hosted H2O and OH in the sample; this is referred to as the calculated H2O-OH. The calculated H2O-OH term assumes all the S in the rock occurs in sulfide minerals-a reasonable assumption for the district samples (see App. 3).

SWIR spectral imaging and feldspar staining
Rock slabs were analyzed using a Specim SisuROCK imaging spectrometer at the University of Alberta, which acquires reflectance data in the SWIR wavelength (l) range of 1,000 to 2,500 nm, at a variable spatial resolution of 0.2 to 1 mm/pixel (App. 2). Absorption positions were retrieved from the second derivative of interpolated spectra, as described in Lypaczewski and Rivard (2018). The term "white mica" as used herein encompasses the end-member minerals and compositional varieties between muscovite, aluminoceladonite, phengite, and illite. In this study, the l absorption position of the whitemica Al-OH feature is classified as short (2,190-2,198 nm), intermediate (2,198-2,204 nm), and long (2,204-2,214 nm); see Appendix 2 for more details. This classification is based on the response of the white mica in fracture halos or the background response if no discrete vein halo is present. Prehnite was identified from the SWIR spectra based on a distinct absorption feature appearing at 1,476 nm, and kaolinite is identified by the presence of an absorption around 2,160 nm.
Rock slabs were etched with hydrofluoric acid and stained with sodium cobaltinitrite to detect K-feldspar and with amaranth to test for plagioclase, calcite, and prehnite. Additional information on the interpretation of staining results is provided in Appendix 2.

MER diagrams
MERs are used to reduce the effects of closure and cast lithogeochemical data in terms that can be related to variations in mineral formulae and chemical reaction stoichiometries and to test for material transfer (Madeisky and Stanley, 1993;Stanley, 2011). More specifically, Pearce element ratio (PER) diagrams use a denominator element that did not participate in material transfer processes. This creates a simple proportionality with material transfer (Madeisky and Stanley, 1993;Stanley and Madeisky, 1994). As a result, it is critical to test whether the chosen denominator is conserved. Figure 3 plots the immobile elements Al versus Ti and Zr for the least altered and altered Guichon Creek batholith rocks. Samples from an originally homogeneous protolith will form a linear array that passes through the origin if the plotted elements are truly immobile, and disparate protoliths will form additional linear arrays and trend lines (Barrett and McLean, 1999;Bushchette and Piercey, 2016). Based on Figure 3, Al, Ti, and Zr are immobile with respect to alteration (i.e., each linear array has a homogeneous protolith). However, there is a variation between different sample suites in terms of Ti, Zr, and several other incompatible elements due to magmatic differentiation in the Guichon Creek batholith (D'Angelo et al., 2017;. Therefore, we cannot apply PER plots sensu stricto to understand magmatic fractionation (Russell and Stanley, 1990). We will demonstrate, however, that by grouping and plotting the data by rock types (defined from a combination of mapping, mineralogical and textural criteria, and Al-Ti compositions) Ti can be used in the denominator in PER plots to elucidate alteration processes.

Assessment of material transfer (element gains and losses)
To assess and visualize element gains, losses, and immobility in select sample pairs, isocon diagrams were constructed (Grant, 1986(Grant, , 2005. Element gains and losses for each alteration facies, based on median values in the data populations, were calculated using equation (1) on a weight basis following Grant's approach (Grant, 1986;Warren et al., 2007) and using Al2O3 as the immobile element (immobility demonstrated below): where CP and CA are the concentrations of the immobile element in the protolith (P) and altered (A) sample, respectively, and ∆C denotes the gain or loss in grams per 100 g of rock for major elements, in parts per million for trace elements, and in parts per billion for Au, Ag, and Hg. Common protoliths for the Guichon Creek batholith rock facies were determined on a sample-by-sample basis using the mineralogical, textural, and Al-Ti composition criteria of McMillan (1985a) and D'Angelo et al. (2017).

Least altered rock (protolith) compositions
The least altered samples (Table 2) were identified using spectral and feldspar staining results and defined by the absence of macroscopic veins, halos, or aplite dikes. These samples have primary plagioclase, K-feldspar, and hornblende with a fresh appearance. The Guichon Creek batholith formed from two magmatic pulses that fractionated from high to low Ti at a relatively constant Al content (Fig. 4A;D'Angelo et al., 2017;Whalen et al., 2017). The decrease in Ti corresponds to a decrease in Mg number (Fig. 4B) and an increase in Si (Fig. 4C). On a plot of 2Ca + Na + K/Ti versus Al/Ti (molar) the mineralogical control on the composition of a sample is indicated by the slope of the line made between the origin and the sample (Madeisky and Stanley, 1993). For feldspar-controlled compositions the slope of the plot is 1, 0.3 if muscovite, and 0 for  chlorite and kaolinite (Stanley and Madeisky, 1994). By plotting the (2Ca + Na + K)/Ti versus Al/Ti slope value against the Al/Ti value, the initial Ca, Na, and K content can be evaluated with respect to magmatic evolution (Fig. 4D). Guichon Creek batholith rocks evolved from high (2Ca + Na + K)/Al values toward compositions controlled by feldspar (Fig. 4D). Least altered Border, Guichon, and Chataway rocks have (2Ca + Na + K)/Al values >1 because of high Ca concentrations (Table 2) associated with hornblende content in excess of 5% (McMillan, 1985a). Unaltered Bethlehem granodiorite has (2Ca + Na + K)/Al between ~1.1 and 1, whereas the Skeena and Bethsaida granodiorites cluster between 1 and 0.95 (Fig. 4D). There is no evidence of primary carbonate (Table 2; D'Angelo, 2016).

Vein and alteration facies characteristics and mapped distribution
Lithogeochemical samples in the Guichon Creek batholith were cataloged into district-scale alteration facies that cap-tured common vein and halo mineralogy based on the feldspar staining and the spectral response from rock slabs (Table  1; Fig 5). The alteration facies outlined below provide the framework for assessing and plotting the lithogeochemical data in subsequent sections. Potassic: K-feldspar: This facies comprises veinlets and discontinuous fracture fills of quartz and chalcopyrite with 1-to 2-cm K-feldspar halos (Fig. 5A). The highest density (~5 cm/m) of veinlets with K-feldspar halos is located over the Valley deposit. The presence of K-feldspar halos continues, at a lower density, for ~2 km west of known mineralization; beyond this distance they are absent (Figs. 2, 5A). A second K-feldspar halo domain occurs over Lornex and Highmont that extends asymmetrically up to 1 km to the northeast in narrow structures and ~2 km to the southeast ( Fig. 2A).
Sodic-calcic: Epidote-(albite): Epidote veins with K-feldspar-destructive albite halos compose the sodic-calcic facies The median oxide and elemental values for the least altered (protolith) and alteration facies populations are presented; data shown correspond to analysis used in elemental gain and loss calculations and diagrams 1 Calculated fields based on S, C, and LOI analysis and on the assumption that all CO2 is associated with CaCO3, and S as sulfide species (Fig. 5B). Vein margins are typically irregular and diffuse (Table 1). In albite-bearing halos primary mafic minerals are replaced by chlorite with accessory titanite and lesser actinolite. At Valley-Lornex and Highmont, albite alteration is crosscut by bornite-bearing, quartz-coarse-muscovite veins (Table 1). Calcite veinlets crosscut and refractured epidote-(albite) facies veins (Fig. 6A, B). Around the Valley-Lornex deposits in the Bethsaida unit, the epidote-(albite) facies is crosscut by prehnite and calcite veinlets and is locally overprinted by prehnite-(white-mica-prehnite) propylitic alteration of varying intensity ( Table 1).
Domains of epidote veins (>1.25 cm/m) with K-feldspardestructive albite halos have a nonconcentric distribution in the Guichon Creek batholith ( Fig. 2A; Table 1). Wallrock alteration is generally restricted to within 2 to 3 cm of epidote veins or fractures and occurs in fracture-controlled sodic-calcite domains that extend up to 7 km away from the porphyry centers (Fig. 2). Some isolated domains of sodic-calcic facies occur in mafic Border rocks. If these are excluded the total mapped area of sodic-calcic facies is ~34 km 2 .
White mica-chlorite: Quartz-(muscovite) and fine-grained white-mica-chlorite alteration: Quartz-(muscovite) facies comprise coarse-grained (~0.3-1.5 cm) gray muscovite, with a short-wavelength Al-OH feature, which typically occurs in the halos of quartz ± chalcopyrite and bornite veins (e.g., Fig. 5C; Table 1). Pervasive, selectively pervasive, and fracture-controlled fine-grained (<1 mm), white mica with predominantly intermediate-wavelength Al-OH features constitute the white-mica-chlorite alteration assemblage (Table 1; Fig. 5D). Veinlets of Cu-Fe sulfide, and locally specularite, are intimately associated with white-mica-chlorite facies (Table 1). In this paper, quartz-(muscovite) and white-mica-chlorite facies are grouped together as whitemica-chlorite facies because of their mineralogical similarity and close spatial association with the mineralization in the porphyry Cu centers (Table 1). Calcite most commonly occurs as microveinlets with white-mica-chlorite facies but is also locally intergrown with fine-grained white mica, chlorite, and Cu-Fe-sulfide in halos (Figs. 6C, D). Additionally, calcite formed irregular blebs in well mineralized late-stage Cu-rich veins and in white mica and clay-altered fault zones that crosscut main stage Cu mineralization at the porphyry centers (Table 1).
The quartz-(muscovite) facies occurs with the main-stage Cu mineralization at Valley-Lornex and locally at Highmont (Table 1; Fig. 2). White-mica-chlorite ± carbonate alteration occurs with sulfide mineralization at Bethlehem, J.A., and Highmont. It also overprints quartz-(muscovite) vein and halo domains at Valley and Lornex (Table 1; Fig. 2). The distribution of the white-mica-chlorite facies generally is limited to the sulfide-bearing domains within the porphyry Cu centers but is also recognized in narrow (1-10 m) zones at several regional Cu occurrences (Table 1; Fig. 2). Variable amounts of kaolinite and montmorillonite (i.e., intermediate argillic alteration; Table 1) typically overprint relict feldspars outside of the white-mica-chlorite alteration halos and zones at the porphyry centers. Overall, however, kaolinite and montmorillonite are primarily restricted to the porphyry Cu centers and some Cu occurrences.
Propylitic: Epidote-(K-feldspar): This subfacies comprises epidote veins with weakly developed and patchy K-feldspar ±  the plagioclase sites; epidote-(albite), sodic-calcic facies. Note the whitening of the feldspars and the absence of the yellow stain that defines the K-feldspar-destructive albite halo around the epidote vein fill. Trace, short-l white mica and prehnite also present in the halo. C. Quartz ± bornite veins with diagnostic coarse-grained, gray, muscovite halos; quartz-(muscovite) subfacies of white-mica-chlorite facies. Pervasive trace kaolinite alteration of feldspar results in ineffective staining of the feldspar sites. The coarse muscovite is characterized by short-to intermediate-l spectral response. D. Pervasive fine-grained white mica-chlorite; white-mica-chlorite facies. K-feldspar is mostly stable, but plagioclase is pervasively altered to fine-grained white mica with intermediate-l spectral response. Mafic minerals are selectively altered to chlorite. Note the late prehnite veinlets (pink stained). E. Epidote vein with K-feldspar halo; epidote-(K-feldspar) subfacies of propylitic facies. Plagioclase is partially altered to K-feldspar within the halo, whereas mafic minerals are selectively and pervasively altered to chlorite. Note the prehnite veinlet with plagioclase-destructive short-l white-mica and prehnite halo, which possibly crosscuts the epidote veinlet. F. Prehnite ± pumpellyite veinlet with fine-grained white-mica-prehnite halo; prehnite-(white-mica-prehnite) subfacies of propylitic facies. Note the prehnite veinlet pink stains and the reddening within the alteration halo. K-feldspar is stable, whereas plagioclase has been selectively and incompletely replaced by short-l white mica and accessory prehnite (small pink-stained grains within the halo). Mineral abbreviations: ab = albite, bn = bornite, ca = calcite, chl = chlorite, cp = chalcopyrite, ep = epidote, kfs = K-feldspar, musc = muscovite, plag = plagioclase, prh = prehnite, qz = quartz, wm = white mica.

K-Feldspar
The prehnite-(white-mica-prehnite) facies is the most common and widely distributed alteration in the batholith. The highest density of this vein type is centered around the porphyry Cu centers ( Fig. 2A), but veins also occur out to the margins of the batholith (cf. Fig. 2B).

PER, MER, and mobile elements characteristics of altered rocks
The altered samples were organized by host rock to reduce the effects of fractionation-related chemical trends (i.e., Fig.  4) and to improve visualization of the results. The samples are colored by the alteration facies and any overprint present and symbolized hierarchically based on the presence of (1) whitemica-chlorite halos, (2) albite halos, (3) K-feldspar halos, and (4) prehnite-(white-mica-prehnite) halos (Figs. 7-9).
Potassic: K-feldspar: Potassically altered rocks plot on the feldspar control line and overlap with other alteration facies in the (2[Ca-C] + Na + K)/Ti versus Al/Ti diagram (Fig. 7A). In terms of K/Ti versus (Al-2[Ca-C]-Na)/Ti the K-feldspar-altered rocks plot close to or on the K-feldspar control line (Fig.  7B); the data that plot high on the K-feldspar control line contain more K-feldspar than points lower on the K-feldspar control line. Many samples containing K-feldspar alteration overlap with the least altered rock field, but a subset shows K-  Figure 6C. E. Calcite occluding an epidote vein with a prehnite-fine-grained white-mica halo associated with prehnite-(prehnite-white-mica) alteration. Note the turbidity of the feldspar in plane-polarized light (PPL) is due to porosity and fine-grained secondary mineral inclusions. F. Hornblende phenocryst replaced by chlorite, calcite, and epidote associated with the prehnite-(prehnite-white-mica) alteration facies. Mineral abbreviations: ab = albite, bn = bornite, ca = calcite, chl = chlorite, cp = chalcopyrite, ep = epidote, f.g. = fine-grained, kfs = K-feldspar, musc = muscovite, plag = plagioclase, prh = prehnite, qz = quartz, tit = titanite, wm = white mica. feldspar addition. Potassically altered samples that have been overprinted by a white-mica-chlorite assemblage are shifted toward the muscovite control line (Fig. 7B). A few K-feldsparaltered rocks hosted by Guichon and Chataway granodiorite plot above the least altered rock domain and have high Rb/Sr and low Rb/Ba (Fig. 8). In the other host rocks, however, the K-feldspar facies is indistinguishable from the least altered and prehnite-(white-mica-prehnite) facies rocks in terms of Sr-Ba-Rb (Fig. 8). The K-feldspar facies, however, can be discriminated based on elevated Cu concentrations (i.e., >100 ppm in Bethsaida) and, generally, has high K/Th values compared to the least altered rocks (Fig. 9). Sodic-calcic: Epidote-(albite): Epidote-(albite)-altered rocks mostly plot on the feldspar control line overlapping with the least altered rocks (Fig. 7A). In the Bethsaida granodiorite, however, numerous albite-altered samples are shifted toward the muscovite control line (Fig. 7A). Samples altered by albite plot below the least altered domains on or close to the K-feldspar control line (Fig. 7B). Sodiccalcic facies rocks that have been overprinted by white micachlorite plot on the muscovite control lines, but at low K/Ti values (blue-colored X symbol, Fig. 7A, B). The variation in the proportions of K-feldspar, muscovite, and kaolinite in the epidote-(albite) facies in Bethsaida granodiorite at low K/Ti values is evident in Figure 7B. On a plot of K/(Al-2[Ca-C]) versus Na/(Al-2[Ca-C]), i.e., with compositions projected from anorthite, the epidote-(albite)-altered Bethsaida granodiorite is shifted toward the albite   K / (Al-2(Ca-C)) C node relative to the least altered rock field (Fig. 7C). Approximately 30% of the epidote-(albite)-altered samples in Bethsaida granodiorite exhibit Na addition, and numerous samples of this facies are situated on the albite-muscovite tie line (Fig. 7C). Albite-altered rocks form arrays in Sr-Ba-Rb space that extend from least altered host-rock compositions toward high Sr/Ba and low Rb/Sr (Fig. 8). Similarly, albitealtered rocks that were overprinted by white mica-chlorite plot in the same field as the epidote-(albite) facies. The high Sr/Ba response corresponds to samples with low K/Al and high Na/Al. Epidote-(albite) facies rocks exhibit a wide range of K/Th that partially overlaps with the least altered rocks, but overall they have lower K/Th and Cu contents (Fig. 9).
White mica-chlorite: Samples identified as white-micachlorite-altered plot between the feldspar and muscovite control lines, where the slope of the line between the datum and the origin corresponds directly to the amount of white mica in the sample (Fig. 7A). The white-mica-chlorite facies rocks have a wide range of K/Ti values that plot along the muscovite control line (Fig. 7B); samples with high K/Ti on the control line are mostly quartz-(muscovite) veins and halos from the Valley pit. There appears to be substantial K, Ca, and Na transfer associated with some white-micaprehnite alteration. Rocks affected by white-mica-prehnite alteration have higher Rb/Sr and higher Rb/Ba than least altered and prehnite-(white-mica-prehnite) altered rocks (Fig. 8). This facies can be further distinguished by elevated K/Th (Fig. 9).
Propylitic: Epidote-(K-feldspar) and prehnite-(white micaprehnite): Propylitically altered rocks predominantly plot on or close to the feldspar control line (Fig. 7A). The compositional changes associated with the propylitic subfacies are within the range of the natural variability for K, Na, and Ca, and background white-mica content in each protolith ( Fig. 7; Table 2). In terms of Rb-Sr-Ba concentrations, the propylitic subfacies are mostly indistinguishable from the other alteration facies (Fig. 8), though a few samples have high Sr/Ba and low Rb/Sr. Similarly, the propylitic subfacies cannot be discriminated from the least altered rocks based on K/Th value (Fig. 9). A few samples of prehnite-(white-mica-prehnite) alteration, however, do have elevated Cu concentrations and high K/Th relative to the least altered domain and sodic-calcic-altered rocks (Fig. 9).

Volatile (H2O-OH and CO2) characteristics of altered rocks
By plotting the molar ratios of Ca/Ti against CO2/Ti, the composition of carbonate present in the whole-rock samples can be estimated (Fig. 10A). Most Guichon Creek batholith samples plot close to the anorthite, prehnite, and epidote trend line, but some extend toward, and form an array along, the calcite control line. Samples with a slope between vertical Fig. 8. Strontium-Ba-Rb ternary diagrams for Guichon and Chataway, Bethlehem and Skeena, and Bethsaida host rocks. Albite-altered rocks show a shift toward high Sr/Ba and Sr/Rb, whereas rocks with white-mica-chlorite alteration have high Rb/Ba and low Sr/Rb. Prehnite-(prehnite-white mica) and least altered rocks, and most K-feldspar-altered rocks, plot in the same location within the Sr-Ba-Rb ternary system, per host-rock group. Albite-altered rocks overprinted by white mica-chlorite have a range of values but appear to preserve a high Sr/Ba signature. 2) Presence of K-feldspar-destructive albite halo 3) Presence of K-feldspar halo and 1 contain some carbonate that is interpreted to be calcite; this is supported by electron probe microanalysis results (App. 2).
The δ 18 O and δ 13 C values of carbonate from 30 wholerock pulps are reported in Table 3. Four least altered rock samples contained carbonate below the detection limits of the analytical technique. Carbonate from altered rocks in the Guichon Creek batholith has δ 18 O values between 9.5 and 19.2‰ with an average value of 13.1‰, and δ 13 C values that range from -7.5 to -1.6‰, with an average value of -4‰. No systematic variation in carbonate isotope composition with respect to alteration facies is evident (Table 3).

Material transfer
Isocon diagrams of altered and fresh granodiorite have Al2O3, TiO2, P2O5, and Zr concentrations with constant ratios and Fig. 9. Copper concentrations versus K/Th for Guichon and Chataway, Bethlehem and Skeena, and Bethsaida host rocks. The epidote-(K-feldspar) alteration facies is predominantly associated with the more mafic host rocks (Guichon and Chataway) and is not elevated in Cu or K/Th relative to least altered rocks, and prehnite-(white-mica-prehnite) facies. K-feldspar alteration facies in Bethsaida partially overlaps with the least altered and prehnite-(prehnite-white-mica) field but some samples extend toward higher K/Th and correspond to samples with high K/Ti in Figure 7B. Additionally, rocks with K-feldspar halos and no epidote vein fill are distinguished by elevated Cu contents. White-mica-chlorite-altered samples have high K/Th and generally higher Cu than the other facies, the inverse is true for epidote-(albite) facies in each host rock. Guichon and Chataway Cu (ppm) Color legend Epidote-(albite) overprinted by prehnite-(white-mica-prehnite) Epidote-(albite) overprinted by white-mica-chlorite Biotite overprinted by epidote-(albite) Epidote-(K-feldspar) overprinted by prehnite-(white-mica-prehnite) K-feldspar K-feldspar overprinted by prehnite-(white-mica-prehnite) K-feldspar overprinted by white-mica-chlorite White-mica-chlorite White-mica-chlorite overprinted by prehnite-(white-mica-prehnite) Prehnite-(white-mica-prehnite) least altered rocks that form a line that projects back to the origin, indicating that these elements were immobile during alteration (Fig. 11;App. 4). Zirconium concentrations are more variable in the mafic host rocks than in the felsic units, i.e., they are less conserved with respect to magmatic processes (Table 2). Additionally, the concentration of TiO2 (and mafic minerals) in the least altered host rocks is low (<1%) and more variable than the concentration of Al2O3 (Table 2). Consequently, we elected to use Al2O3 in material transfer (element gain and loss) calculations, consistent with the methods outlined by Grant (1986). Here, the material transfer of the major oxides is expressed as a percent change from the least altered rock median values, and the other elements are in parts per million. Material changes greater than 1 standard deviation of the least altered rock suite are considered to be significant (Table  4). Figure 12 shows the median response for the major oxides, volatiles, and select minor elements per alteration facies for all samples hosted by the Guichon (mafic) and Bethsaida (felsic) granodiorites (Table 4).
In the Guichon granodiorite isocon plot, the sodic-calcic alteration is highly depleted in Fe2O3, K2O, Ag, Au, Ba, Rb, and Li, is moderately depleted in Cu, Co, Fe2O3, Ni, and Pb, and has a minor depletion in MgO, Ce, and Sr concentrations (Fig.  11A). Conversely, Na2O, CaO, Mn, Zn, and H2O-OH are rich in the sodic-calcic-altered sample relative to the least altered values (Fig. 11A). The element changes in sodic-calcic alteration hosted in Guichon granodiorite shown on the isocon plot (Fig. 11A) are confirmed by the strong depletion in K2O (-32%) and Ba (-260 ppm) and enrichments in CaO (11%), Na2O (4%), and Mn (30 ppm) in the Bethsaida granodiorite shown in Figure 12. More significant enrichment in CaO (18.5%) and Na2O (8.2%) locally occurred with sodic-calcic alteration, as evidence by the 75 th percentile values of the sample suite (Table 4; Fig. 12). The mafic host rock, the Guichon granodio- rite, is more depleted in Fe2O3 (-9%), Ba (-440 ppm), and Cu (-65 ppm) and possibly Zn (25 th percentile = -11 ppm) relative to concentrations of these elements in more felsic Bethsaida granodiorite. No significant changes in SiO2 concentrations occurred with epidote-(albite) alteration. Both host rocks show significant addition of CO2 (533-643%) and H2O-OH (74-89%) during sodic-calcic alteration (Fig. 12). Potassic alteration in Bethsaida granodiorite is enriched in K2O (15%) and depleted in CaO (-6%), but the latter is within 1 standard deviation of the least altered sample suite. Concomitant depletion of CaO with enrichment in K2O, however, is indicated by the 25 th percentile value of CaO change in the potassically altered sample suite (Table 4). Depletion of CaO is more pronounced in the Guichon than in the Bethsaida granodiorite (Fig. 12). K-feldspar-altered rocks are strongly enriched in copper (Fig 12), and depletion in Mn may be occurring in Bethsaida granodiorite (see next section), but generally it is within the natural variation for this host rock. The potassic facies is not significantly enriched in CO2 and H2O-OH.
No significant changes in the major oxides are associated with either of the subfacies of propylitic alteration, but both are enriched in H2O-OH (41-56%) relative to the least altered sample suite. A wide range of enrichment in CO2 characterizes the prehnite-(white-mica-prehnite) propylitic subfacies (Table 4).

Spatial distribution of alteration intensity and material transfer
In order to compare the intensity of the feldspar-destructive white-mica alteration between the different district-scale host rocks in the Guichon Creek batholith, the (2[Ca-C] + Na + K)/Al value for each sample was normalized to its least altered equivalent host rock. The delta values (Δ) between the least altered host rock (2Ca + Na + K)/Al value (Fig. 4D) and 1 (the feldspar control line) were subtracted from the measured (2[Ca-C] + Na + K)/Al value for each sample.
The porphyry Cu centers exhibit moderate to intense white-mica alteration (i.e., (2[Ca-C] + Na + K)/Al values of  (Figs. 13, 14) and CaO. Mineralization at the Valley deposit is characterized by >30% increase in K2O relative to the least altered Bethsaida granodiorite, whereas the response at Lornex and Highmont is more mixed, possibly due to the presence of albite alteration. Sodic-calcic domains are primarily characterized by depletion in K2O with some addition of Na2O (Fig. 13) ± CaO. A ~1-km-wide domain west and southwest of the Valley Cu center, with and without white-mica formation, has locally elevated K2O concentrations (and SiO2 to a lesser degree). This domain is also weakly to moderately enriched in Cu and appears to be locally depleted in Mn (Fig. 14). For approximately 5 km south of Lornex and Highmont, depletion in K2O is recorded in some samples within the mapped sodic-calcic domain, and weak white-mica alteration occurs proximal to interpreted structures, but overall the footprint in this area is heterogeneous. Potassium addition also occurred in the rocks south of Lornex at the southernmost fringes of the mapped sodic-calcic domain. White-mica alteration, without SiO2 but with Mn addition, can be discerned in rocks up to 1 km northeast of Bethlehem (Figs. 13, 14). Transfer of K2O and Cu to the host rock and low values of (2[Ca-C] + Na + K)/Al are found at several regional Cu occurrences (Figs. 13, 14). The cluster of samples depleted in K2O west of the Kathleen Cu occurrence (Fig. 13) do not appear to be albite altered (based on feldspar staining) and are interpreted to be a false negative caused by lower initial K2O concentrations in the host rock in this area compared to the least altered median value used in the mass transfer calculations.
The addition of CO2 during alteration occurs throughout the batholith, but the highest density of rocks significantly enriched in CO2 is within ~3 km of the porphyry centers (Fig.  15). Similarly, a higher proportion of samples enriched in H2O-OH occurs near the porphyry centers (Fig. 15). An alternative method to investigate enrichment in H2O-OH is to use Mg (molar) as a denominator; this has the effect of normalizing the data to its mafic mineral content, because mafic minerals are the primary source of water and hydroxyl in least altered rocks. Combining a H2O-OH/Mg-normalized term with the change in CO2 concentration enhances the pattern of volatile enrichment. For ~6 km south of Lornex, 2 km southwest of Valley, 0.5 km northeast of Bethlehem, and to a lesser degree 1 km northeast of Highmont, rocks are variably enriched in H2O-OH and CO2 (Fig. 15). These areas mostly overlap with the domains of mapped propylitic and sodic-calcic alteration around the porphyry centers.

Host-rock control on altered rock mineralogy and composition
The ability to predict alteration products and chemical changes during fluid-rock interaction is important for rec- Fig. 11. Isocon diagrams with selected and weighted elements in which the protolith versus the altered samples are plotted (after Grant, 1986). A. Epidote-(albite)-altered Guichon (sample 2015KB60 shown). B. White-micachlorite facies Skeena (sample 2014GL173 in Fig. 5C). Elements plotting above the isocon line are poor in the altered rock, whereas elements below the line are rich relative to the protolith.   Table 4). Similarly, the removal of Cu ± Zn, presumably by chlorite alteration of Cu-bearing primary hornblende, is more pronounced in the mafic host rocks (Tables 2, 4; Fig. 12). These phenomena are interpreted to be primarily caused by a larger degree of chemical disequilibrium between the composition of the mafic host rocks and the fluids causing the alteration than between the more felsic rocks and the fluid. Hydrothermal alteration mineralogy is largely dependent on the host-rock bulk composition and the water-rock ratio during exchange (Berger and Velde,Fig. 12. Material transfer diagrams for the main alteration facies in the Guichon and Bethsaida intrusive phases. Major oxide gains and loss expressed as % change from the least altered rock suite median value. Minor and trace element gains and loss expressed as ppm. Abbreviations: ep = epidote, kfs = K-feldspar, prh = prehnite, wm = white mica. Least altered rock ± 1σ     Reed, 1997); while the latter can only be inferred, the host-rock composition can be estimated. Accordingly, in the Guichon Creek batholith, epidote > prehnite veins with propylitic halos are more abundant in the hornblendedominated mafic host rocks (3.3-5.4 wt % Fe2O3; Table 2), rather than in the Fe-poor (2.3 wt % Fe2O3; Table 2) and biotite-dominated Bethsaida granodiorite; here, prehnite > epidote veinlets are more common with propylitic alteration (Fig. 2). Prehnite has been rarely reported in propylitically altered rocks around porphyry Cu deposits. Two notable exceptions documented propylitic alteration of basaltic host rocks in the Cadia district (New South Wales, Australia, Wilson et al., 2003) and in andesitic host rocks at Collahuasi (northern Chile, Djouka-Fonkwe et al., 2012). Epidote most commonly forms at temperatures ranging from 200° to 350°C during propylitic alteration (Bowman et al., 1987;Norman et al., 1991). Prehnite and epidote can coexist over a limited temperature range around 250°C; however, the stability field of prehnite expands with decreasing temperature to ~150°C, below which zeolites predominate (Bird et al., 1984;Digel and Gordon, 1995;Bird and Spieler, 2004). In the temperature range of 200° to 350°C, the most important control on the stability of epidote and prehnite in the system CaO-Al2O3-Fe2O3-SiO2-H2O-HCl is the activity of Ca, Fe, and Si, although high activities of CO2 in the fluid can also inhibit prehnite and epidote formation (Wheeler et al., 2001;Bird and Spieler, 2004). At a fixed temperature, the epidote stability field is larger at higher activities of Fe and lower activities of Ca and Si relative to prehnite (Bird et al., 1984;Digel and Gordon, 1995). On a mineral-stability basis and in rocks with similar Ca content, therefore, epidote formation is favored in a more Ferich wall rock. The observations from the Guichon Creek batholith suggests that in Fe-poor and Ca-rich plutonic or volcanic rocks, prehnite veinlets can occur in propylitic alteration zones. Additionally, in the case of epidote-(albite) alteration, the degree of metasomatism may be limited in Fe-and K-poor rocks relative to a host rock that is rich in these elements.

Material transfer and fluid sources
The chemistry of most K-feldspar-altered samples in the Guichon, Chataway, Bethlehem, and Skeena host rocks overlaps with the least altered rock domains, indicating only minor amounts of K were added during alteration (e.g., Fig.  6). A few K-feldspar-altered samples hosted in the Bethsaida unit plot above the least altered rock domain. Around the porphyry Cu centers, generally the material transfer during potassic alteration is minor but can be locally characterized by lower CaO and Mn and higher K2O, K/Th, and Cu (± Ag) concentrations compared to the least altered rocks. The white-mica-chlorite facies are strongly enriched in K2O and Cu (and other ore and pathfinder elements; Fig. 12; Table 4), variably enriched in SiO2, and depleted in Na2O and CaO in and around the porphyry centers. The material transfer caused by potassic and white-mica-chlo-  rite alteration is consistent with formation of the alteration assemblages from predominantly magmatic-hydrothermal fluids interacting with the wall rocks (Ulrich and Heinrich, 2001;Sillitoe, 2010).
The sodic-calcic alteration in the Guichon Creek batholith is represented by veins of epidote with albite ± fine-grained white-mica halos principally characterized by depletion in K2O with varying degrees of Na2O and CaO addition. Ahmed et al. (2019) also noted that the albite alteration in the Yerington district did not consistently occur with Na addition but was typically associated with >50% CaO addition.
A key criterion for the formation of sodic-calcic alteration is the presence of Cl-bearing fluid coupled with a prograding thermal path, which can result in an exchange reaction between primary K-feldspar and secondary albite (Carten, 1986;: KAlSi3O8 + Na + (aq) = NaAlSi3O8 + K + (aq). ( This reaction is consistent with the mineralogical changes demonstrated by the PER plots and the K loss and localized Na gain during sodic-calcic alteration in the Guichon Creek batholith. Chloritization of biotite can also result in K loss from the host rock. The removal of Fe, Cu, Ag, and Zn during albitization would be facilitated by the heating of a Cl-bearing fluid, because the solubility of these metals increases with increasing temperature (Hemley and Hunt, 1992). In this study we have also shown that Ba and Rb are removed from the rock during albitization, and these elements are inferred to be primarily derived from igneous K-feldspar (Heier, 1962). Based on our data, however, we can only speculate on the fate of the elements mobilized by the fluid during sodic-calcic alteration (i.e., K, Fe, Cu, Ba, and Rb); perhaps they are redeposited in the porphyry center during mixing with magmatic-hydrothermal fluids.
There does not appear to be any significant material transfer of CaO or K2O during epidote-(K-feldspar) propylitic alteration in the Highland Valley Copper district. The alteration is primarily characterized by the addition of H2O-OH (10-93%) ± CO2 (3-465%; Table 4). The K required for the patchy replacement of plagioclase by K-feldspar in vein halos of this facies is interpreted to be locally sourced from biotite during chlorite alteration. Additionally, in order to stabilize secondary K-feldspar over white mica, more alkaline conditions are required for epidote-(K-feldspar) alteration relative to the prehnite-(white-mica-prehnite) subfacies (Hemley and Hunt, 1992). Epidote-(K-feldspar) veins are primarily hosted in Chataway or Guichon granodiorites, and their formation may be related to contact metamorphism and hydrothermal circulation caused by intrusion of the second pulse of magma in the Guichon Creek batholith.
Alternatively, the white mica formed in equation (3) is mostly paragonite, in which case little addition of K is required, and the Na liberated from plagioclase is consumed. The hydration of hornblende and formation of chlorite liberates Ca: Ca2Mg4Al(Al0.5Si3O11)2(OH)2 + Mg 2+ (aq) +2H + + 10H2O = Mg5Al(AlSi3O10)(OH)8 + 2Ca 2+ (aq) + 4H4SiO4(aq). (5) The Fe, Mg, Ca, and Si in aqueous solution, derived from fluid-mineral interactions (eq. 3-5), is inferred to be primarily consumed by the formation of prehnite and epidote fracture fills, as there is no significant material transfer of these elements associated with this alteration facies (e.g., Figs. 6, 12). The source of Al for these minerals is unclear, but lab experiments have shown that albitization can locally mobilize Al and Ti (trace levels) from plagioclase via interface-coupled dissolutionreprecipitation at the micron scale (Hövelmann et al., 2010). Additionally, turbidity in feldspars, ubiquitous in the propylitic and sodic-calcic halos in the Guichon Creek batholith, indicates porosity formation (Plümper and Putnis, 2009;Putnis, 2015). Thus, the Al required for minerals precipitated in fractures, as opposed to simply replacing wall rock, is inferred to be locally sourced from feldspars during alteration and concomitant porosity generation in the halo. Nevertheless, it should be noted that at the Guichon Creek batholith, rock compositions do indicate that Al is immobile at the lithogeochemical sample scale (e.g., Figs. 3, 13). Although not accounted for in mass balance calculations, propylitic alteration also caused oxidation of primary magnetite to variable mixtures of hematite and ilmenite and localized reddening within the alteration halo.
Calcite has a retrograde solubility with temperature at fixed pressure, and its precipitation is driven by an increase in fluid pH and CO2 degassing (e.g., eq. 6), as opposed to fluid cooling (Rimstidt, 1997). Carbonate δ 18 O and δ 13 C values of about 13 and 4‰, respectively, suggest that the CO2 fixed in the rocks as calcite is magmatic in origin (Kyser, 1986;Taylor, 1986;Djouka-Fonkwe et al., 2012). Calcite alteration occurs in some rocks associated with white-mica-chlorite facies but mostly manifests as fracture fills. Hydrolysis reactions of feldspar that forms abundant white mica and quartz consume H + ions and, thus, promote calcite precipitation: 2H2CO3(aq) + Ca 2+ (aq) = CaCO3 + CO2 + H2O + 2H + . (6) This reaction is a plausible mechanism for calcite formation in the white-mica-chlorite facies. However, it is less likely to form calcite in veins in the peripheral parts of the porphyry Cu systems (with lower white-mica content). Phase separation of CO2 from a late-stage magmatic volatile phase, on the other hand, is more consistent with the paragenesis of calcite in the altered rocks around the Highland Valley Copper porphyry Cu centers. In this case, the Ca required for the formation of calcite in equation (6) could come from the Ca liberated during chloritization of hornblende in propylitic and/or epidote-(albite) domains, or from Ca removed from Cu-mineralized zones associated with white-mica-chlorite facies. Phase separation of CO2 and calcite formation from a magmatic volatile phase would be enhanced by decompression during a transition from early, high-temperature alteration phases (potassic, sodic-calcic, and early halo-type muscovite), to the lower-temperature and later alteration phases (white-mica-chlorite and propylitic).
Propylitic and sodic-calcic alteration can form from a variety of fluid sources. A common model for the formation of propylitic alteration involves wall-rock reaction with inwardflowing, thermally prograding, meteoric fluids at low waterrock ratio (Meyer and Hemley, 1967;Norman et al., 1991;Reed, 1997). Alternatively, meteoric water with a nonmagmatic brine component (Bowman et al., 1987) or the lateral expulsion of cooling spent magmatic-hydrothermal fluids can generate propylitic alteration peripheral to Cu-bearing potassic or sericite-chlorite alteration domains (Pacey et al., 2020). Sodic-calcic alteration may be caused by the flow of external, hypersaline formation waters  or by seawater-sourced fluids (Orovan et al., 2018;Byrne et al., 2020), heated during inflow into the magmatic cupola region along the margins of potassic alteration and porphyry stocks (Dilles et al., 2000). Propylitic alteration around the Highland Valley Copper porphyry centers is interpreted to have formed from a CO2 degassing, cooling, and spent magmatic volatile phase or from the circulation of an external fluid (or a mixture of both) at moderate water-rock ratios. The overlap of the early sodic-calcic domains with a high density of younger, prehnite-(white-mica-prehnite) veins and alteration ( Fig. 2A) and the presence of cogenetic magmatic volatile-derived calcite in the younger alteration suggest that both seawater-derived fluids and magmatic fluids were present peripheral to the Cu centers at Highland Valley Copper at different times (see Byrne et al., 2020). In other porphyry systems, some elements lost from potassic zones (e.g., Ca, Fe, Co) are redistributed outward, producing metasomatic propylitic alteration domains, primarily through the formation of epidote replacement zones (Pacey et al., 2016;Ahmed et al., 2019); thus, a spent and cooling magmatic fluid could conceivably cause localized Ca metasomatism in propylitic domains but perhaps not deplete the rock in K, Fe, and Cu over broad areas.
In summary, this study benefited from a significant number of lithogeochemical samples (850), which allowed us to assess the district-scale mineralogical and geochemical changes. The presented workflow of linking the feldspar staining and SWIR spectral results to the lithogeochemical analyses at Highland Valley Copper shows that the extent of mineralogical change and material transfer is variable for each alteration facies (i.e., Table 4). This variability can be interpreted to be related to a combination of process such as fluid focusing, total fluid flux (water-rock ratio), and the physiochemical gradients between wall rock and fluid. Additionally, our standardized sampling method, i.e., inclusion of the fresh rock around veins and halo (e.g., Fig. 13A), resulted in lower percent element change relative to least altered host rocks than what would have been achieved if only vein and halo material was included in the sample. As such, we believe the outlined methods will be useful for other studies of, and exploration programs focused on, porphyry Cu systems.

Implications and Conclusions
The workflow presented here-i.e., linking feldspar staining and SWIR spectral results to lithogeochemical analyses-revealed several features of interest: (1) staining highlights weak and cryptic K-feldspar destruction that did not cause sufficient material transfer to be confidently distinguished from protolith compositions; (2) K-feldspar alteration west of the Valley deposit is easily detectable with staining, but generally only caused weak alkali metasomatism; however, this facies is characterized by some SiO2 addition and higher-than-background Cu and Ag concentrations; (3) SWIR spectral imaging distinguished white mica in samples with an intermediate-to long-wavelength Al-OH feature that is closely related to magmatic fluids and proximal parts of the footprint (see Alva-Jimenez et al., 2020); and (4) staining and spectral imaging identified a subfacies of propylitic alteration, epidote-(K-feldspar), which is probably related to the second magmatic pulse in the Guichon Creek batholith and implies a long history of propylitic (green rock) alteration in the district prior to porphyry Cu mineralization.
We demonstrate that MER and PER diagrams are an effective way to assess subtle mineral changes and material transfer caused by alteration processes but are dependent on correct protolith assignment. Accordingly, careful protolith discrimination using a combination of mapping, textural and mineralogical criteria, and paired immobile elements characteristics (e.g., Al versus Ti, Sc versus Ti, Zr, Th, and Nb) is required. LOI and C and S data can be used to estimate rock H 2O-OH and carbonate mineral concentrations. Furthermore, these data can be easily combined with major oxide determinations to model carbonate (e.g., Fig 9A) and, in some circumstances, sulfide mineral species within the porphyry footprint.
Disparate field and analytical data sets were integrated to identify mappable domains of alteration facies, to quantify material transfer and mineralogical changes, and to detect the total extent of the footprint at Highland Valley Copper. A large area of weak fracture-controlled K-feldspar-dominated potassic alteration developed around the porphyry stocks and barren quartz vein domain (Table 1) and is locally characterized by minor gains in SiO2, K2O, and Cu-Ag (Fig. 16, time 1). Localized depletion in K2O, Fe2O3, and Cu and enrichment in Na occur in structurally controlled sodic-calcic alteration domains that formed a large (~34 km 2 ) nonconcentric footprint outboard of well-mineralized and K-rich zones (Fig. 16, times 2, 3). The sodiccalcic alteration is interpreted to be caused by the heating and inward flow of seawater-derived fluid that mixed with magmatic volatile phases at some of the porphyry centers . Well-developed feldspar-destructive white-mica alteration, indicated by (2[Ca-C] + Na + K)/Al values <0.85, addition of SiO2, K2O, and Cu-Ag, and de-pletion in CaO and Na2O, occurs with quartz-(coarse muscovite) and white-mica-chlorite alteration in the Cubearing parts of the porphyry system (Fig. 16, times 2, 3). Outflow of magmatic volatiles from the porphyry centers, coupled with CO2 degassing, caused incipient calcite precipitation and peripheral white-mica-chlorite alteration with an intermediate to long l Al-OH absorption position (Fig. 16, time 4). An admixture of externally derived wall-rock equilibrated fluid and CO2-degassing magmatic volatile phase depleted in ore elements is inferred to have generated propylitic alteration over a large area (~60 km 2 ) peripheral to the K-rich and Cu-mineralized domains (Fig. 16, time 5). Although interpreted to have formed from seawater-derived fluid, sodic-calcic-altered rocks received a late-magmatic volatile phase flux that is recorded as a series of crosscutting calcite veinlets that is related to porphyry Cu genesis.
Several observations from this study are transferable to other porphyry systems and have exploration implications. Protolith composition influences alteration mineralogy. In calciumrich and Fe-poor host rocks, prehnite veins with propylitic halos can occur; this may be an underappreciated feature of the footprint in some porphyry systems. If present, sodic-calcic alteration can deplete the protolith in Fe and magnetite concentrations, causing a change in petrophysical properties (e.g., Byrne et al., 2019) that may be detectable using geophysical surveys. At Highland Valley Copper we found that water and magmatic CO 2-rich rocks form the largest coherent lithogeochemical footprint (~60 km 2 ) that extends beyond domains of enrichment in ore and pathfinder elements and pronounced alkali metasomatism. Recognition of large areas of H2O and magmatic CO2-rich rocks in prospective porphyry Cu districts is useful because they indicate the presence of a hydrothermal system, can reduce areas of interest for followup target development, help focus mineral chemistry tools (i.e., Baker et al., 2020;Cooke et al., 2020), and potentially provide a vector toward porphyry Cu centers.