The Paleoproterozoic Falun Zn-Pb-Cu-(Au-Ag) pyritic sulfide deposit in the Bergslagen ore district, Sweden, is enveloped by hydrothermally altered rocks metamorphosed to the lower amphibolite facies. Immobile-element ratios suggest that the alteration precursors were volcanic rocks of mainly rhyolitic to dacitic composition. Least altered examples of these rocks plot along magmatic fractionation trends outlined by late- to post-ore feldspar-phyric metadacite dikes and post-ore granitoid plutons, consistent with a comagmatic relationship between these calc-alkaline, coeval (<10-m.y.) suites. Dolomite or calcite marble, as well as diopside-hedenbergite or tremolite skarn, form subordinate but important lithologic components in the hydrothermally altered zone. Marble occurs as fragments in the massive pyritic sulfide mineralization, suggesting that at least some mineralization formed by carbonate replacement.

Mass-change calculations suggest that the hydrothermally altered volcanic rocks gained Mg and Fe and generally lost Ca, K, and Na. Proximal, quartz-anthophyllite-rich altered rocks additionally gained Si, whereas several types of biotite-rich altered rocks lost this element. These mass changes along with mineral chemical data for anthophyllite, biotite, cordierite, and garnet, and the common occurrence of quartz indicate that chloritization, sericitization, and silicification were the dominant premetamorphic alteration styles. A zonation from distal sericitized and silicified volcanic rocks to intermediate sericitized rocks, partly overprinted by chloritization (Mg-rich chlorite), and proximal siliceous and intensely chloritized (Fe-rich chlorite) rocks has been identified. Furthermore, mass changes in more peripheral parts of the altered zone toward the southeast of the deposit suggest that the alteration weakens gradationally toward the volcanic and subvolcanic rocks surrounding the deposit. These patterns represent vectors toward mineralization.

Intensely chloritized rocks, largely represented by a single, rhyolitic precursor, envelop the central pyritic massive sulfide bodies to the east, south, and west, supporting a structural model in which the massive sulfide mineralization formed the stratigraphically highest preserved unit in the center, surrounded in a tubular manner by stratigraphic footwall rocks. The northern side represents a portion of the footwall, which was separated by a major shear zone. These spatial relationships also have implications for near-mine exploration, since quartz-rich footwall rocks locally host disseminated to semimassive stockwork Cu-Au mineralization.

Cooling of a hot (300°–400°C), acidic (pH ≤4) and reducing fluid carrying metals and sulfur is suggested for formation of stockwork Cu-Au vein mineralization and hydrothermal alteration in the stratigraphic footwall. The Zn-Pb-Cu-rich massive sulfide mineralization is inferred to have formed by fluid neutralization upon interaction with carbonates and mixing with cooler seawater upon fluid entry into porous pumice breccia in a subseafloor setting. Dissolution processes, primary porosity in the pumice breccia, and secondary porosity produced during synvolcanic faulting are all suggested to have contributed to the creation of space necessary for the formation of the massive sulfide mineralization. Falun differs from other deposits of the same type in Bergslagen mainly in the high pyrite content of the massive sulfide mineralization, the absence of related Fe oxide deposits, as well as the dominant replacement of volcaniclastic sediments compared to carbonates. The types of host rocks, the inferred premetamorphic feldspar-destructive alteration types, and the style of mineralization and alteration zonation at the deposit are reminiscent of pyritic volcanogenic massive sulfide (VMS) deposits. However, the importance of chemical trapping by fluid-limestone interaction, as well as the spatial association with subordinate skarn alteration constitute important differences to a classic VMS model.

The Bergslagen ore district in south-central Sweden hosts several thousand Fe oxide as well as base and precious metal sulfide deposits; the latter typically associated with metamorphosed hydrothermally altered zones. The district belongs to the 1.9 to 1.8 Ga Bergslagen lithotectonic unit, which forms a major component in the southwestern part of the 2.0 to 1.8 Ga Svecokarelian orogen within the Fennoscandian Shield, Sweden (Fig. 1A). A major pyritic Zn-Pb-Cu-(Au-Ag) sulfide deposit is situated in the town of Falun in the northwestern part of this ore district (Fig. 1A). Falun has a mining tradition extending over several centuries that terminated in 1992, giving way to its current status as an UNESCO world heritage site. In total, approximately 28 to 35 million metric tons (Mt) of ore were produced at grades of 0.5 to 5% Zn, 0.1 to 1.7% Pb, 0.7 to 4% Cu, 13 to 35 g/t Ag, and 0.5 to 4 g/t Au, varying between different mineralization types (Tegengren, 1924; Grip, 1978; Allen et al., 1996). Despite historically being the type locality for the dominant type of base metal sulfide deposits in the district (Magnusson, 1950, 1953; Allen et al., 1996), relatively little modern research has been conducted at Falun. Genetic models for the deposit, established during the previous century (Geijer, 1917; Koark, 1962), are in need of revision. More recent studies have either a more regional focus (Kresten, 1985; Bromley-Challenor, 1988; Smeds, 1994), or investigated specific aspects of the Falun deposit (Weijermars, 1987; Gavelin, 1989; Åberg and Fallick, 1993).

Several studies have led to a classification of the base metal sulfide deposits in the Bergslagen ore district into two types, Falun belonging to the so-called “Falun” (Magnusson, 1950, 1953) or “SVALS” (strata-bound volcanic-associated limestone-skarn-hosted) type (Allen et al., 1996). Genetic models for this type of deposit involve leaching of metals from mainly submarine felsic volcanic rocks by modified seawater, but vary in process from volcanic-exhalative to subsurface carbonate replacement (e.g., Koark, 1962; Boström et al., 1979; Baker and DeGroot, 1983; Lagerblad and Gorbatschev, 1985; Oen, 1987; Allen et al., 1996; Jansson et al., 2013). Many of these authors have inferred processes for SVALS deposits similar to those at volcanogenic massive sulfide (VMS) deposits, although Allen et al. (1996) showed that the paleoenvironment in Bergslagen differed from typical VMS districts in the predominance of shallow marine volcanism.

A new structural model for the Falun deposit, involving sheath folding during the Svecokarelian orogeny (Kampmann et al., 2016b), in combination with new geochronological data from the deposit, indicating a short time interval (< 10 m.y.) between felsic volcanism and mineralization (Kampmann et al., 2016a), provide the framework for this study. The current work is focused on the character of the hydrothermal system and ore formation at the Falun deposit with the following aims: (1) determine the nature, zonation, and intensity of hydrothermal alteration associated with Falun by mass-change calculations; (2) identify premetamorphic alteration mineral assemblages; and (3) constrain the geologic conditions during the mineralizing event, which have implications for ore formation, later structural modification, and exploration.

Regional geologic framework

Base and precious metal sulfide deposits as well as Fe oxide deposits in the Bergslagen lithotectonic unit are hosted by a folded belt of metamorphosed, predominantly submarine, felsic volcanic and subvolcanic rocks, the emplacement of which was preceded and followed by sedimentation of siliciclastic sedimentary rocks (Fig. 1A). The timing of igneous crystallization of the volcanic and subvolcanic rock suite is constrained between U-Pb (zircon) ages of ca. 1.91 and 1.89 Ga (Lundström et al., 1998; Andersson et al., 2006; Stephens et al., 2009). Allen et al. (1996) suggested that the volcanic succession was deposited in a back-arc basin on continental crust.

Base metal sulfide deposits are generally hosted by the upper parts of the volcanic stratigraphy in spatial association with carbonate rock and skarn (Allen et al., 1996; Stephens et al., 2009). Hydrothermal alteration of volcanic rock and limestone, associated with mineralization, occurred in zones around the base and precious metal sulfide deposits (e.g., Trägårdh, 1991; Allen et al., 1996; Stephens et al., 2009), including Falun (Wolter and Seifert, 1984). The mineralogy of the rocks in these hydrothermally altered zones was modified during the Svecokarelian orogeny. Such metamorphic rocks are termed “altered rocks” in this paper and are named according to their current metamorphic mineralogy. For clarity, the term “altered silicate-rich rock” in this contribution collectively refers to rocks of an inferred volcanic precursor and not the marbles and skarns with an inferred limestone precursor.

The mainly supracrustal sequence was intruded by different suites of plutonic rocks, which form large volumes in the Bergslagen lithotectonic unit (Fig. 1A). These suites have been distinguished on the basis of composition, age, meta-morphic character, and structural relationships (Stephens et al., 2009). The oldest 1.90 to 1.87 Ga GDG intrusive rock suite, represented by granitoid rock, diorite, and gabbro, showing a calc-alkaline trend is the most prominent (Fig. 1A).

Regional metamorphism under low-pressure, greenschist-to amphibolite- and locally granulite-facies conditions, as well as polyphase ductile deformation during at least two cyclic reoccurrences of transpressive deformation (D1 and D2) at around 1.87 to 1.86 and 1.84 to 1.82 Ga affected the Bergslagen lithotectonic unit (Hermansson et al., 2008; Stephens et al., 2009; Beunk and Kuipers, 2012; Stephens and Andersson, 2015; Kampmann et al., 2016b). Transpression was in each case preceded and followed by extension or transtension. The supracrustal rocks and the GDG intrusive rock suite are generally older than all the ductile transpressive deformation and metamorphism in the Bergslagen lithotectonic unit and, thus, modified by these processes. The shift from extensional or transtensional to transpressional tectonics has been attributed to a transfer from retreating to advancing arc modes along a convergent, active continental margin (Hermansson et al., 2008; Stephens et al., 2009; Stephens and Andersson, 2015).

Local geologic setting

The Falun base metal sulfide deposit occurs within a geologic unit with an east-northeast-west-southwest trend, referred to here as the Falun inlier (Fig. 1B), which consists mainly of metamorphosed felsic volcanic and subvolcanic rocks affected by regional alkali alteration that modified the original K and Na contents (Bromley-Challenor, 1988). Mafic dikes are present throughout the inlier, which is surrounded by GDG plutonic rocks and isolated intrusive bodies belonging to younger plutonic suites (Fig. 1B). SIMS U-Pb (zircon) ages of the volcanic and subvolcanic rocks, the felsic dikes, and the GDG plutonic rocks overlap in age within their respective analytical uncertainties at around 1894 Ma, indicating hydrothermal alteration and sulfide mineralization within a narrow time span of < 10 m.y. when intense magmatic activity and burial of the supracrustal rocks prevailed (Kampmann et al., 2016a).

Volcaniclastic rocks in the Falun inlier include fine-grained, volcanic silt- to sandstone and pumice breccia (Table 1). Volcanic silt- to sandstone is typically equigranular, although feldspar-phyric varieties occur. Lamination and bedding of varying thickness are common. Pumice breccia is common in the volcanic units close to the Falun deposit and has been observed in road cuts within <2 km of the deposit (Fig. 2). It is texturally and structurally similar to the volcanic silt- to sandstone, but exhibits varying contents of elongate wisps of biotite up to several centimeters in length (Fig. 3A). These biotite wisps are interpreted to represent compacted pumice fragments, derived from a proximal volcanic source (McPhie et al., 1993; Allen et al., 1996; Lasskogen, 2010). The amount of pumice fragments in this facies type varies from pumicebearing (up to 10 vol %) types to massive pumiceous rocks. The occurrence of compacted pumice fragments and associated volcanic facies types are well documented in volcanic rocks of the Bergslagen lithotectonic unit (e.g., Allen et al., 1996).

The volcaniclastic rocks can be distinguished from finegrained, feldspar- to quartz-feldspar-phyric rocks of similar composition (Fig. 3B, Table 1), interpreted on the basis of their homogeneity throughout the inlier and the presence of resorption embayments in the quartz phenocrysts, as coherent subvolcanic intrusive rocks. Undifferentiated volcanic or subvolcanic rocks are usually massive (nongraded and nonbedded), fine-grained (aphanitic and aphyric), and cannot be assigned to any facies type.

Feldspar-phyric metadacite dikes (Fig. 3C, Table 1) crosscut the sulfide mineralization and the surrounding hydrothermally altered zone (Fig. 2; Koark, 1962; Kampmann et al., 2016b). Along their margins, amphibolite with mafic dike precursors is common, containing mainly hornblende, plagioclase, and minor magnetite; the two types of dikes apparently showing a composite character (Kampmann et al., 2016b).

The amphibolites also form isolated intrusions elsewhere in the Falun inlier. The field relationships and crystallization ages suggest emplacement during a late stage of, or after, the hydrothermal alteration event, but prior to metamorphism and deformation (Kampmann et al., 2016a).

Continuous rapid burial down to the plutonic regime in the < 10-m.y. time span resulted in the premetamorphic intrusion of the area by medium-grained granitoid rocks (Fig. 3D), and subordinate microgranite, belonging to the GDG intrusive rock suite (Fig. 1B, Table 1). Intrusions of mafic to intermediate, probably GDG, rocks, which were affected by large-scale F2 folding, occur both inside and surrounding the inlier (Fig. 1B, Table 1).

Polyphase ductile deformation modified the volcanic and subvolcanic rocks, as well as the dikes and the GDG plutonic rocks, while associated metamorphism peaked at 2.5 ± 1 kbars and 550° ± 50°C, i.e., indicating low-pressure and lower amphibolite-facies conditions (Wolter and Seifert, 1984). At least one planar fabric is usually defined by minor, elongate biotite and/or amphibole needles. The inlier is dominated by F2 major folds with S asymmetry and axial surface traces trending east-west to northeast-southwest (Fig. 1B), which deform an earlier S1 planar fabric. D2 ductile shear zones with the same strike and a steeply plunging mineral stretching lineation are also present. The pyritic massive sulfide mineralization at Falun is located directly south of one such shear zone and regional mapping work by the Geological Survey of Sweden has identified another with west-northwest-eastsoutheast strike immediately to the south of the deposit (Figs. 1B, 2). Three-dimensional modeling of the deposit revealed steeply SSE-plunging, rod-shaped, and downward converging mineralized bodies down to ~450 m below sea level, related to F2 structures (Weijermars, 1987) inferred to be sheath folds (Fig. 4; Kampmann et al., 2016b).

Field mapping, core logging, and sampling

Field studies were carried out in the open pit, along road cuts, and in small outcrops close (< 2 km) to the Falun deposit. Mapping in the open pit was limited by the nonaccessibility of the medium- and high-level benches. A total of 66 samples for lithogeochemistry were acquired during field work. Sixty-seven additional samples were collected from drill cores logged at the Geological Survey of Sweden’s national archive in Malå, Sweden (46 samples) and at Drake Resources Ltd. in Falun (21 samples). Characterization of the samples was aided by thin section microscopy, with 12 thin sections selected for mineral chemistry analysis.

Fifty-two lithogeochemistry samples, acquired from five drill cores from the area southeast of the Falun deposit (referred to as Falun SE; Fig. 2) from Lasskogen (2010) are incorporated here. Six samples in this study from a road cut are also included in the Falun SE sample group (Table A1). This area is located in an intermediate position between intensely altered rocks at the Falun deposit and the volcanic and subvolcanic rocks farther away from the deposit inside the Falun inlier.

In the course of the recently completed regional mapping project by the Geological Survey of Sweden focused on the Falun inlier, an additional 50 samples, representing rocks inside and immediately surrounding the inlier, were collected and analyzed for lithogeochemistry, and are included here. Generally, the sampling of felsic volcanic or subvolcanic rocks, as well as intrusive rocks in the regional mapping project was performed with the intention to analyze rocks unaffected by hydrothermal alteration.

Three-dimensional modeling

In order to establish the spatial relationship between rock units in Falun SE and the WNW-ESE-striking shear zone south of the deposit (Fig. 2), the positions of five drill cores extending below Falun SE (Lasskogen, 2010) were calculated with trigonometric functions using spatial information included in legacy mine data, made available to us by the Geological Survey of Sweden. A three-dimensional geologic model of these drill cores, including their logs (Lasskogen, 2010), and the shear zone was completed using the software GOCAD 2009.3 (Paradigm Ltd.), including the SPARSE-plugin (Mira Geosciences Ltd). Dip and strike values for the shear zone are based on measurements taken along a road cut where a portion of this zone is exposed (Fig. 2). This model complements a previously published three-dimensional model of the more proximal parts of the Falun deposit (Kampmann et al., 2016b), where a detailed description of the three-dimensional modeling workflow used here was presented.

Lithogeochemistry

The samples acquired for this study were cleaned and weathered parts at the surface removed with a rock saw. Further sample preparation, involving crushing and pulverizing using low-chrome steel grinding mills, was carried out by ALS Minerals prior to analysis at their laboratory. Lithogeochemical analysis was conducted using Lithium Metaborate Fusion and inductively coupled plasma atomic emission spectroscopy (ICP-AES) for major oxides, and inductively coupled plasma mass spectrometry (ICP-MS) for resistive trace elements, including rare earth elements (REE). For analysis of C and S, the LECO combustion furnace infrared detection technique was used. Volatile components were measured by loss on ignition (LOI). Base metals, including Ag and gold-related trace elements, were analyzed by ICP-MS after a four-acid (HNO3- HClO4-HF-HCl) or an aqua regia digestion, respectively. If returned values for Ag, Cu, Mo, Ni, Pb, and Zn were above the upper detection limit for the methods described above (>10 000 ppm), these samples were additionally analyzed by four-acid digestion ore-grade ICP-AES.

For quality control, three Geological Survey of Sweden inhouse tonalite standards were randomly interspersed with the study samples. Additionally, five study samples representing different rock types were sent to ALS Minerals a second time for duplicate analysis, as well as to Acme Laboratories for analysis using the same techniques. The returned values for major oxides and immobile elements were, in all cases, identical to the original results within 5% uncertainties of each other. There was a discrepancy regarding the Zr mass fraction of 7% for one sample between the two laboratories, the ALS duplicate analysis returning a nearly identical value (< 1σ difference) to the original analysis. Somewhat higher divergences between the two laboratories were observed regarding LOI analysis, with differences up to 35% of each other. Litho-geochemical data are visualized below in diagrams using the software ioGAS™. All binary element-element graphs show element mass fractions on a volatile-free basis. The methods and quality control used for lithogeochemical analysis of drill core samples from Falun SE were reported in Lasskogen (2010).

Mineral chemistry

Mineral chemistry analyses of anthophyllite, biotite, cordierite, and garnet from the different altered silicate-rich rocks in the hydrothermally altered zone at Falun were performed on polished thin sections. A Jeol JXA-8530F HyperProbe electron probe microanalyzer (EMP) at the Department of Earth Sciences, Uppsala University, Sweden, was used, with standard operating conditions of 15 kV (accelerating voltage) and 15 nA (beam current). Calibration and standardization were carried out using international standard reference material.

Based on earlier work (Koark, 1962), and recent mapping by the Geological Survey of Sweden and Kampmann et al. (2016b), the hydrothermally altered zone at Falun extends several 100 m from the central pyritic massive sulfide mineralization and covers a surface area of at least 5 km2 (Fig. 1B). Outside this zone, the volcanic rocks appear to be variably (subtly to moderately) sericitized and silicified. This alteration commonly intensifies in the vicinity of the Falun deposit or one of the minor base metal sulfide deposits in the area (Fig. 1B). The hydrothermally altered zone at the Falun deposit is dominated by different intensely altered silicate-rich rocks with variable contents of quartz, biotite, cordierite, garnet, anthophyllite, and retrogressive chlorite and talc (Koark, 1962; Kampmann et al., 2016b). A gradation from fine-grained, undifferentiated volcanic rock into altered silicate-rich rock and, ultimately, to disseminated to semimassive Cu-Au mineralization was mapped along cores drilled on the eastern side of the deposit (Fig. 2). These observations support the interpretation that the volcanic and subvolcanic rock suite is the precursor to the altered silicate-rich rocks at the Falun deposit (Törnebohm, 1893; Högbom, 1910; Geijer, 1917; Koark, 1962).

The two most common types of altered silicate-rich rock are biotite-quartz-cordierite-(anthophyllite) schist (BQC schist; Fig. 5A, Table 1) and quartz-anthophyllite-(biotite-cordierite/almandine) metamorphic rock (QA rock; Fig. 5B, Table 1). Biotite-quartz-cordierite-(anthophyllite) schist is prominent north of the ductile shear zone forming the northern boundary of the pyritic massive sulfides, whereas QA rock surrounds the central massive sulfide body to the south of the shear zone (Fig. 2). Subordinate types of altered silicate-rich rocks are biotite-almandine-(anthophyllite) and biotite-cordierite- (anthophyllite) schists (BA and BC schist, respectively; Table 1), as well as talc-chlorite-(quartz-biotite) schist (TC schist; Table 1). The TC schist is restricted to the ductile shear zone at the northern boundary of the massive sulfide body and in a smaller shear zone along part of the south-western boundary (Fig. 2). Chlorite and talc pseudomorphs after cordierite and anthophyllite, respectively, are present in this rock type (Weijermars, 1987; Kampmann et al., 2016b). Chlorite is common as a retrogressive mineral in most types of altered silicaterich rocks at Falun, staurolite has been observed in accessory amounts associated with cordierite (Wolter and Seifert, 1984). Andalusite, gahnite, and gedrite are accessory phases in QA rock.

Dolomite or calcite marble, as well as diopside-hedenbergite or tremolite skarn are locally present in the hydrothermally altered zone (Geijer, 1917; Koark, 1962; Kampmann et al., 2016b; Fig. 2). Feldspar-phyric metadacite dikes were emplaced in the hydrothermally altered zone, but do not exhibit the intense alteration described above; only a weak to moderate sericitization and silicification appears to be present.

The different types of mineralization at the Falun deposit include the following:

  1. Disseminated to semimassive, vein-style Cu-Au mineralization (chalcopyrite, pyrite and local pyrrhotite and sphalerite) in quartz-rich QA rock (Fig. 5C), also hosting auriferous quartz veins on the eastern side of the deposit (Åberg and Fallick, 1993) and local cassiterite. Gold mineralization in association with Cu mineralization grade between 0.5 and 4 g/t Au (Tegengren, 1924; Allen et al., 1996).

  2. Pyritic massive sulfide that forms a major central body, with Zn-Pb-Cu-rich sulfide (sphalerite, galena, chalcopyrite) mineralization in its center and minor, more Cu-rich zones along its margins (Kampmann et al., 2016b; Figs. 2, 5D-F). Pyrrhotite locally occurs instead of pyrite in the massive sulfides.

  3. Disseminated sulfides (pyrite and chalcopyrite) in TC schist along ductile shear zones (Fig. 5G).

  4. Zn-Pb-(Ag) mineralization in subordinate bands, pods, and lenses of massive diopside-hedenbergite or tremolite skarn (Fig. 2; Geijer, 1917; Koark, 1962; Lasskogen, 2010; Kampmann et al., 2016b).

The major part of the central pyritic massive sulfide body contains either quartz (Fig. 5D) and minor anthophyllite, biotite, and cordierite as gangue minerals or lacks nonsulfide gangue minerals (Fig. 5E). The silicate gangue is interpreted to represent altered relics of the felsic host unit; quartz may in this case represent an alteration product or preserved and recrystallized quartz phenocrysts. In the outer part, toward the contact with the altered silicate-rich rocks, rounded fragments of dolomite or calcite marble varying in size between 1 mm and several meters occur in the massive sulfide mineralization (Fig. 5F; Geijer, 1917). The marble fragments locally contain amphibole, which indicates a possible mixture with volcanic material during the carbonate sedimentation. Veins of ophicalcite (serpentine and calcite) and pyrite crosscut larger fragments of dolomite marble (Fig. 5H) at the ground surface in the northwestern part of the central pyritic massive sulfide body (Fig. 2). The northwest-southeast elongation of the massive sulfides appears structurally discordant to the general eastwest to northeast-southwest trend of rock contacts surrounding this mineralization type. However, where visible, the contacts between massive sulfides and altered silicate-rich rocks are commonly sheared and transposed into the main ENE-WSW-trending foliation (Fig. 2; Kampmann et al., 2016b).

Subsequent polyphase ductile deformation and metamorphism resulted in the development of planar and linear tectonic fabrics in the altered silicate-rich rocks, the intensity of which is dependent on the relative content of quartz and mica (Kampmann et al., 2016b). The inferred sheath fold structural model for the deposit, together with the presence of the same types of intensely altered silicate-rich rocks on all sides of the pyritic massive sulfides (Fig. 2) implies that a stratigraphic hanging wall to the Falun deposit is absent in the direct vicinity of the deposit (Kampmann et al., 2016b).

In drill cores from Falun SE, Lasskogen (2010) noted an abrupt transition southward and structurally upward from intensely altered silicate-rich rocks, marble, and skarn, to apparently unaltered pumice breccia and volcanic silt- to sandstone, therein interpreted to represent lithologies of the stratigraphic hanging wall. Three-dimensional modeling of these drill cores and the west-northwest-east-southeast shear zone in the area predicts that this zone coincides spatially with the boundary between altered and apparently unaltered rocks, assuming a steep dip for the zone (Fig. 4). The model is supported by field observations along a road cut in Falun SE, where a steeply S-dipping shear zone with predominantly north-sideup displacement (Fig. 3E) separates QA rock including Cu mineralization to the north, from amphibolite and metamorphosed volcanic silt- to sandstone and pumice breccia (Fig. 3F) to the south. It is inferred that the units on either side of the shear zone have been tectonically decoupled, thereby raising doubts around the stratigraphic interpretation in the drill cores (Lasskogen, 2010), which includes a correlation of rock units on both sides of this shear zone.

Lithogeochemical data and approach used in their interpretation

Median and median absolute deviation (MAD) values for the lithogeochemistry results from each sampled rock type are presented in Table 2. Arithmetic and geometric means and standard deviations are herein avoided, since histogram analyses suggest a skewed data distribution for the majority of elements and, for this reason, normal or lognormal distribution cannot be assumed (Reimann and Filzmoser, 2000). The five different types of altered silicate-rich rocks are referred to using their respective abbreviations, indicating their dominant mineral associations. In order to obtain representative statistics, values below detection limit were converted to half the detection limit prior to the calculation of median and MAD values (e.g., Verbovšek, 2011). For an overview of all analytical results, the reader is referred to the electronic supplementary material (Table A1).

The approach to the lithogeochemical data in this study is based on the division of the rocks at and around the Falun deposit into two main groups; volcanic, subvolcanic, and altered silicate-rich rocks on the one hand; and dikes and granitoid rock, diorite, and gabbro intrusive rocks on the other. A third, subordinate group of samples comprises marbles and skarns. The workflow follows, in general terms, the approach to hydrothermal alteration in volcanic districts described by MacLean and Barrett (1993) and Barrett and MacLean (1994, 1999). Immobile elements were used herein to identify precursors to altered rocks, evaluate the magmatic relationship between the different suites, and calculate mass changes resulting from alteration. This workflow has been successfully applied in, for example, a study of hydrothermal alteration and chemostratigraphy at the metamorphosed Kristineberg VMS deposit, situated in the Skellefte ore district in northern Sweden (Barrett et al., 2005). The immobile character of Zr, Al2O3, and TiO2 has been demonstrated at some base metal sulfide deposits in the Bergslagen ore district (e.g., Jansson et al., 2013; Jansson and Allen, 2015).

Characterization of hydrothermal alteration and identification of least altered samples

Volcanic, subvolcanic, and altered silicate-rich rocks: A total of 47 volcanic and subvolcanic rocks from the Falun inlier were sampled in the field (including Falun SE) and from drill core, both at the deposit and, using Lasskogen (2010), in Falun SE (Fig. 4). Classified after volcanic facies, the samples comprise 13 felsic volcanic silt- to sandstones, 13 felsic pumice breccias, eight felsic subvolcanic rocks, and 13 undifferentiated volcanic rocks.

The field classification of the volcanic and subvolcanic rocks as subalkaline rhyolite to dacite is confirmed by a plot of Nb/Y versus Zr/TiO2 for rock classification (Winchester and Floyd, 1977; Fig. 6A). The different facies types in this suite all form one cluster in this diagram, the majority plotting as rhyolite and approximately 30% as dacite; two samples show an andesitic composition. The samples overlap the field defined by felsic volcanic and subvolcanic rocks from elsewhere in the Bergslagen lithotectonic unit (Fig. 6A; Stephens et al., 2009). Elevated SiO2 mass fractions (i.e. >70 wt % SiO2; Fig. 6B; Table A1) are not consistent with this classification, but can be explained by the observed silicification in a part of the volcanic and subvolcanic rock samples.

The 106 samples of altered silicate-rich rocks show a similar pattern from mainly rhyolitic to subordinate dacitic and rare andesitic/basaltic compositions in a plot of Nb/Y versus Zr/TiO2 (Fig. 6C). The altered silicate-rich rocks should plot at the same point in this diagram as their respective precursor rocks, if Zr, TiO2, Nb, and Y were immobile during alteration (Fig. 6C). These samples show a large variability in SiO2 mass fraction (ca. 20–85 wt %; Tables 2, A1; Fig. 6D), indicating that intense enrichment or depletion in SiO2 occurred during hydrothermal alteration of the inferred predominantly felsic volcanic precursor rocks.

To assess hydrothermal alteration qualitatively, the volcanic and subvolcanic rocks, as well as the altered silicate-rich rocks are shown in an alteration box plot, designed to evaluate alteration in VMS systems (Fig. 7; Large et al., 2001). The alteration box plot has also been applied successfully to metamorphosed VMS deposits (Theart et al., 2010). The alteration indices used in this diagram do not consider mass changes of Si in the hydrothermal system. The majority of the volcanic and subvolcanic rocks from the Falun inlier lie inside the unaltered rhyolite and dacite fields (Fig. 7A). However, a few samples show trends toward both the albite and sericite end members, suggesting that some sodic alteration or sericitization, respectively, has affected these samples. Volcanic siltto sandstones and pumice breccias from the Falun SE area show a clearer trend toward the sericite end member (Fig. 7A), indicating increased sericitization of volcanic rocks in the vicinity of the Falun deposit. No clear alteration patterns with regard to the different volcanic facies types can be deduced from Figure 7A.

The majority of altered silicate-rich rocks form a tight cluster in the chlorite-pyrite end-member corner of the same alteration box plot (Fig. 7B). Apparently less intensely altered samples show one direct trend to this corner from the least altered fields and a second array from the sericite end member. The trend from least altered volcanic and subvolcanic rocks to sericite (Fig. 7A) joins up with the sericite to chlorite spread for altered samples (Fig. 7B) to form a continuous trend. Since different altered silicate-rich rock types overlap in the diagram, few obvious differences in alteration style or intensity can be inferred from the box plot. Generally, altered silicate-rich rocks from Falun SE and BQC schists show a trend toward the sericite end member, whereas QA rocks tend to plot closer to the chlorite-pyrite end-member corner (Fig. 7B). This is in accordance with the observation that the QA rock is the dominant altered silicate-rich rock type close to the central pyritic massive sulfide body.

To calculate mass changes in hydrothermally altered rocks, it is necessary to identify the least altered samples among the precursor rock suite to the more intensely altered rock types (MacLean and Barrett, 1993). In binary plots of immobile elements (e.g., Zr, TiO2, Al2O3), least altered comagmatic rocks should plot along common magmatic fractionation trends. Net gains or losses of mobile elements during hydrothermal alteration would lead to a decrease or increase, respectively, in the concentrations of immobile elements, but the ratio of one immobile element to another would be retained (MacLean and Barrett, 1993). Thus, rocks sharing the same precursor should plot along a linear array anchored at the origin, where samples plotting above the fractionation trend record net mass loss, whereas those that plot below record net mass gain.

A sample of the volcanic and subvolcanic rock suite at Falun is considered least altered if it fulfills the following criteria:

  1. Groundmass and phenocryst mineralogy are in agreement with the rock classification. Major element mass fractions (Table A1) are within the natural range of the classified rock type when compared with data from Nockolds (1954).

  2. Lack of visible textures, which would be indicative of hydrothermal alteration (Gifkins et al., 2005).

  3. Absence of minerals typical of the altered silicate-rich rocks (i.e., cordierite, anthophyllite, garnet) and no anomalously high contents of biotite and quartz, chlorite and talc.

  4. The plot of Nb/Y versus Zr/TiO2, the TAS diagram (Le Bas et al., 1986; not shown here) and the alteration box plot are in agreement regarding rock classification for the sample.

  5. Other lithogeochemical indicators are below certain threshold values (S <0.2 wt %, LOI <3 wt %). These values are consistent with previous lithogeochemical studies on volcanic rocks in the Bergslagen ore district (Allen et al., 1996; Jansson and Allen, 2015).

Following the criteria adopted above, only three felsic volcanic or subvolcanic samples have been identified as least altered, one pumice breccia (PNY120185A), one subvolcanic rock (CMR120073A), and one undifferentiated, fine-grained volcanic rock (9/13-4). The lithogeochemistry of these samples and their composition, based on their position in the TAS classification diagram (Le Bas et al., 1986), is summarized in Table 3. Binary immobile element plots for all the volcanic and subvolcanic rocks, as well as the altered silicate-rich rocks are shown in Figure 8A-C, and the least altered samples and their highly tentative fractionation trends for different element pairs are highlighted. Alteration trends toward the origin, intersecting the proposed fractionation trend, connect all remaining samples.

Dikes and granitoid rock, diorite, and gabbro intrusive rocks: The base metal mineralization at Falun was intruded by feldspar-phyric dacite dikes, spatially associated with mafic dikes, which are now metamorphosed to metadacite and amphibolite, respectively (Geijer, 1917; Koark, 1962; Kampmann et al., 2016a; Fig. 2). Classification according to the plot of Nb/Y versus Zr/TiO2 (Fig. 6E) is in agreement with the petrographic classification of the feldspar-phyric metadacite dikes (21 samples), based on phenocryst populations of plagioclase > K-feldspar > quartz. However, they have SiO2 mass fractions above 70 wt % (Tables 2, A1; Fig. 6F). Ground-mass sericitization and silicification are present in some metadacite dikes. Si enrichment due to hydrothermal alteration (i.e., silicification) is inferred from high amounts of quartz in the groundmass, which is inconsistent with the phenocryst mineralogy, and the lithogeochemical characteristics. The amphibolite dikes (16 samples) show a basaltic to andesitic composition (Fig. 6E).

Metamorphosed mafic to intermediate and felsic granitoid rock, diorite, and gabbro intrusive rocks show no textural or mineralogical evidence of hydrothermal alteration, with the exception of some sericitization of feldspar in the rocks (Kampmann et al., 2016a). Five mafic to intermediate samples are classified as gabbro to diorite in a TAS plutonic rock classification diagram (Fig. 6G) and the remaining felsic granitoid rock, diorite, and gabbro rocks (18 samples) plot in the granite and granodiorite fields in the same diagram (Fig. 6G), in agreement with the petrographic classification.

The dike and granitoid rock, diorite, and gabbro samples have been plotted in the alteration box plot (Fig. 7C, D, respectively), in order to be able to qualitatively compare the style and intensity of their alteration with the felsic volcanic, subvolcanic, and altered silicate-rich rocks. Amphibolite dikes plot in or close to the least altered andesite-basalt field, whereas the feldspar-phyric metadacite dikes plot either in the least altered dacite or rhyolite fields, or are shifted toward the sericite end member (Fig. 7C). The granitoid rock, diorite, and gabbro rocks plot inside or close to the least altered fields (Fig. 7D). Three samples plot toward the albite endmember corner, indicating some introduction of albite.

Using the same procedure as for the felsic volcanic or subvolcanic rocks and altered silicate-rich rocks described above, only one feldspar-phyric metadacite dike, but the majority of the felsic granitoid rock, diorite, and gabbro samples (11), qualify as least altered. The lithogeochemistry of these samples and their composition, based on their position in respective total alkali-silica (TAS) classification diagrams (Le Bas et al., 1986, Middlemost, 1994), is summarized in Table 3. Most felsic granitoid rock, diorite, and gabbro rock samples define linear arrays in the binary immobile-element graphs involving Zr, Al2O3, and TiO2 (Fig. 8D-F), which are consistent with magmatic fractionation trends (MacLean and Barrett, 1993). However, a few samples plot at some distance to the lines, which may be due to effects of minor mass change (hydrothermal alteration), or to the presence of more than one magmatic fractionation trend among the felsic granitoid rock, diorite, and gabbro rocks. The single least altered dike sample also lies close to the felsic granitoid rock, diorite, and gabbro fractionation trend in each diagram. However, the majority of feldspar-phyric metadacite dikes define a trend from the least altered sample toward the origin, which is especially evident in the Zr versus TiO2 diagram (Fig. 8D). Mass gain due to hydrothermal alteration is inferred.

Precursors to hydrothermally altered rocks

Immobile elements (e.g., Zr, Al2O3, TiO2) can be used to identify possible subsuites among the felsic volcanic and subvolcanic precursor rocks, samples belonging to the same subsuite forming clusters in ratio-ratio diagrams and lines passing through the origin in binary graphs, independent of the effects of hydrothermal alteration (MacLean and Barrett, 1993; Barrett et al., 2005). Three major clusters and one minor cluster have been inferred in one such ratio-ratio diagram (Fig. 9A). The former show well-correlated linear alteration trends in binary graphs, distinctive with significantly different slopes on the Zr versus TiO2 and TiO2 versus Al2O3 diagrams (Fig. 9B, D), and overlapping in the Zr versus Al2O3 graph (Fig. 9C). Samples in the major clusters lie either inside or close to the rhyolite field in the plot of Nb/Y versus Zr/TiO2, or consistently in the dacite or rhyolite field in the same diagram (Fig. 9E). No trend can be identified for the samples in the minor cluster that plot distinctly outside the alteration trends for all the other samples (Fig. 9B-D). They show an andesitic or basaltic composition (Fig. 9E).

The approach above has resulted in the identification of at least two major precursors for altered silicate-rich rocks within the volcanic and subvolcanic rock suite. These precursors (1) and (2) can be distinguished by their predominantly rhyolitic and dacitic compositions, respectively. All except one pumice breccia sample are part of precursors (1) or (2), which also contain the vast majority of altered silicate-rich rocks (Fig. 9A-E). For this reason, pumice breccia is inferred to be the most dominant facies type among the volcanic rocks at the Falun deposit, prior to hydrothermal alteration, consistent with the field observations close to the deposit. Two additional components, referred to as precursors (3) and (4), have a rhyolitic and andesitic to basaltic character, respectively. Precursor (3) comprises samples corresponding to different volcanic facies, including the two volcanic silt- to sandstone samples. It has to be kept in mind that it is possible that even more subsuites exist within the precursors defined above, but there is no basis to justify a further subdivision at present. The spatial distribution of the respective precursors at the Falun deposit is shown in Figure 9F.

Magmatic relationships between coeval igneous suites

The coeval relationship between the three main igneous suites in and around the Falun inlier, the volcanic and subvolcanic rocks, the feldspar-phyric metadacite dikes and the felsic granitoid rock, diorite, and gabbro plutonic rocks (Figs. 1B, 2; Kampmann et al., 2016b), is compatible with but not definitive evidence of a comagmatic relationship. Comagmatic suites should define identical magmatic fractionation trends, the correct identification of which is crucial for calculating mass changes (MacLean and Barrett, 1993; Barrett and MacLean, 1994; Barrett et al., 2005). This can be evaluated by plotting samples of least altered rocks in binary plots involving different immobile-immobile, immobile-incompatible, and immobile-mobile element pairs.

The least altered samples of all three igneous suites together define linear or near-linear trends in the immobile-immobile (Fig. 10A-C) and the immobile-incompatible (Fig. 10D-F) graphs. These trends closely resemble magmatic fractionation trends of felsic rocks elsewhere, both inside (e.g., Roberts et al., 2003; Barrett et al., 2005) and outside (e.g., MacLean and Barrett, 1993; Barrett and MacLean, 1994) the Fennoscandian Shield.

An inferred comagmatic relationship means that the linear pattern for the three suites taken together may serve as a proxy for the fractionation trend of the volcanic and subvolcanic rocks. To further assess this, immobile-mobile element pairs were plotted. Magmatic fractionation trends for several major elements expressed as oxides (e.g., CaO and MgO) should resemble polynomial, concave trend lines with negative slopes (Barrett and MacLean, 1994; Barrett et al., 2005). Least altered samples of all igneous suites at Falun show only moderate correlations (Fig. 10G-J). Four samples (one feldspar-phyric metadacite dike and three felsic granitoid rock, diorite, and gabbro rocks) have slightly lower SiO2 and elevated CaO, MgO, and FeO(total) values, which may indicate that these samples do not follow the same fractionation trend or that mass changes caused by some hydrothermal alteration had occurred. If these four samples are excluded, regression lines with r2 values varying between 0.37 and 0.59 can be calculated for the remaining samples in each immobile-mobile element plot (Fig. 10G-J).

Rare earth element (REE) data normalized to chondrite (Boynton, 1984) for the inferred precursor rock groups of volcanic and subvolcanic rocks, as well as other igneous suites at and around the Falun deposit define negatively inclined slopes (Fig. 11). The light REEs are enriched by approximately one order of magnitude relative to the heavy REEs. A calc-alkaline affinity for all groups and suites is inferred, which is supported by La/Yb ratios consistently around 6 or higher (MacLean and Barrett, 1993). Different slopes and La/Yb ratios for amphibolites and feldspar-phyric metadacite dikes do not support a strictly composite dike relationship between these spatially associated rock types (Kampmann et al., 2016b). Negative Eu anomalies are recognizable in all the felsic rocks (Fig. 11A, C, F, H). Such anomalies are a common primary feature in felsic rocks that have undergone fractional crystallization. However, they can also result from the breakdown of plagioclase feldspar during hydrothermal alteration and this may be an additional cause of these anomalies in the altered silicate-rich rocks (Barrett et al., 2005). Furthermore, the ranges in especially the light REE values of the precursor groups may have been caused by some mobility of the REE in the altered silicate-rich rocks in connection with hydrothermal alteration (MacLean, 1988; Barrett and Sherlock, 1996).

Apart from the positive Eu anomalies, the shapes of the normalized REE curves for the marbles and skarns are similar to the magmatic and altered silicate-rich rocks (Fig. 11). This indicates the presence of a volcanic component in the marbles and skarns, which is further confirmed by the abundance of elements such as Si (median 14.75 wt % SiO2) and Al (median 3.4 wt % Al2O3) in the marble (Table 2).

Magmatic affinity can also be assessed by using binary graphs involving certain immobile and incompatible elements such as Yb, Th, and La (MacLean and Barrett, 1993; Barrett and MacLean, 1994, 1999). A calc-alkaline affinity for the volcanic and subvolcanic rock suite, including the altered silicate-rich rocks as their alteration products, as well as for dikes and granitoid rock, diorite, and gabbro intrusive rocks, is compatible with this suggestion (Fig. 12). However, some samples fall in the transitional and even tholeiitic fields in the Yb versus La graph, especially among the felsic volcanic and subvolcanic, and altered silicate-rich rocks (Fig. 12B), possibly related to some mobility of La.

Mass changes during hydrothermal alteration in different lithologies

The trend lines defined above (Fig. 10) give some support to a comagmatic relationship between the three igneous suites, and can be used as proxies for the magmatic fractionation trends of the volcanic and subvolcanic rocks, in order to calculate mass changes in the altered silicate-rich rocks. The polynomial fractionation trend for Zr versus TiO2 (Fig. 10A) has been selected for the calculation procedure in each suite. Nearly identical results on mass change have been obtained using the linear trend for Zr versus Al2O3, which has a higher correlation coefficient (Fig. 10B). The original Zr content of each sample is provided by the intersection between the magmatic fractionation trend and the individual alteration line connecting the sample with the origin (MacLean and Barrett, 1993; Barrett and MacLean, 1999). These values are then inserted into the fractionation line equations for each major element (e.g., Fig. 10G-J) to obtain precursor values. All the calculated mass changes are given in absolute weight percent for the oxide values. Due to the focus on the felsic range of samples in this study, mafic to intermediate rocks have been excluded from the calculations.

Mass changes in the major and minor elements for each rock type (referred to as ΔSiO2, ΔMgO, etc.) are presented in Table 4 as median and MAD values. The results for ΔFeO(total), ΔMgO, ΔCaO, and ΔK2O are illustrated in binary graphs against ΔSiO2 (Fig. 13, Table 4). Apart from Na (the plot for which is not shown), the remaining major elements show little or no mass change (Table 4), which can be explained by their immobile nature (Ti) or lesser effects on their mass fractions during hydrothermal alteration. The mass changes for Ba and Mn are also insignificant (<0.2 wt % expressed as oxides). Calculations for trace elements have not been carried out due to the difficulty to attain reliable magmatic fractionation trends for these elements. For several elements (e.g., Ga, Nb, Sc, Th, Y, as well as the heavy REE (Tb, Dy, Ho, Er, Tm, Yb, Lu) and Gd), immobility is indicated by plots against Zr.

Since formation of limestone was common at times of reduced volcanism in the Bergslagen lithotectonic unit, it is possible that particularly the fine-grained, silty, reworked volcaniclastic rocks may contain a minor component of sedimentary carbonate material (Allen et al. 1996, 2003). The presence of carbonate in the volcanic, subvolcanic, and altered silicaterich rocks at Falun is tentatively indicated by the occurrence of C in the lithogeochemical data (up to 1.36 wt %, with one outlier at 7.47 wt %, which was not included in mass change calculations; Table A1). Graphite has not been observed in these rocks. Thus, the resulting mass fraction of CaO in such samples may be affected by the presence of carbonate material, which may influence the mass-change calculations that assume solely magmatic CaO.

Mass changes in volcanic and subvolcanic rocks from the Falun inlier, including Falun SE, are variable but commonly below 8 wt % for major elements. Weakly altered felsic volcanic and subvolcanic rocks mostly display a general gain in Si, which can be attributed to the observed silicification of these rocks. The strongest gain in this rock suite is in the two samples of volcanic silt- to sandstones from the area outside Falun SE (median 57.16 wt % ΔSiO2; Table 4), whereas the pumice breccias in the same area gained the least Si (median 5.72 wt % ΔSiO2; Table 4). Both the highest gain in K (median 6.02 wt % ΔK2O) and the highest loss in Na (median –2.21 wt % ΔNa2O) are also observed in the same volcanic silt- to sandstones (Table 4), with the other facies types generally showing minor mass changes regarding these alkali elements. Such mass changes are indicative of sericitization (Barrett and MacLean, 1993). Gains of Na in the subvolcanic rocks (median 1.24 wt % ΔNa2O) may be due to weak albitization of this facies type (Gifkins et al., 2005).

The majority of altered silicate-rich rocks at the Falun deposit (BA, BC, BQC, and TC schists) are depleted in Si and enriched in Mg and Fe (Fig. 13A, B). This is a characteristic feature of chloritization during hydrothermal alteration (e.g., MacLean and Barrett, 1993; Barrett and MacLean, 1994, 1999; Gifkins et al., 2005). In contrast, the QA rock consistently gained Si (up to 138.47 wt % ΔSiO2; Table 4), which is indicative of significant silicification of this rock type and is in accordance with its high quartz content. Depletions in Ca and nearly complete loss of K and Na are apparent in all the altered silicate-rich rocks (Fig. 13C, D; Table 4), which can be explained by the breakdown of feldspar during hydrothermal alteration involving sericitization (e.g., Gifkins et al., 2005).

The predominant QA rock, bordering the central massive sulfide body on its western, southern, and eastern sides, experienced a gain in Fe that exceeded the gain in Mg. By contrast, the BQC schist gained about two times as much Mg relative to Fe, as expressed in median oxide values (Table 4). In hand specimen, this is reflected by the occurrence of almandine garnet in the QA rock, especially in units with little or no cordierite, whereas the BQC schist has a generally higher cordierite content. Positive Mg mass change is more than three times higher than Fe enrichment in the TC schist (Table 4), allowing this schist to be classified also as a dominantly Mgenriched altered silicate-rich rock type. Similar relationships between the elements Fe and Mg are apparent in the BA (dominantly Fe-enriched) and BC (dominantly Mg-enriched) schists, consistent with the dominance of almandine garnet as opposed to cordierite in the former and the occurrence of cordierite instead of garnet in the latter.

Limited mass gains of Si and K, as well as loss of Na, which is near-complete in some samples (Table 4), occurred in the feldspar-phyric metadacite dikes, indicating silicification and sericitization (Fig. 13D, Table 4). The dikes do not show the large gains of Mg and Fe characteristic of the altered silicaterich rocks that experienced chloritization. The contrasting style and intensity of alteration in the feldspar-phyric metadacite dikes supports previous suggestions (Kampmann et al., 2016a) that these dikes were emplaced either during the waning stage of the mineralizing event, being affected by lower temperature hydrothermal fluids in the mineralizing system, or after mineralization and affected by syntectonic fluid movement.

The felsic GDG rocks experienced only minor mass changes (median below 1 wt % for all major oxides except SiO2 with 3.08 wt % gain in median values; Fig. 13; Table 4), reflecting the common absence of alteration texture or mineralogy, or only minor sericitization observed in this rock type. Since the GDG rocks were emplaced after synvolcanic hydrothermal alteration (Kampmann et al., 2016a) and distal to the Falun deposit, mass changes in this rock type can be attributed to syntectonic alteration. Some felsic GDG samples experienced mass changes in Si, K, and Na that exceed the mass changes in some of the volcanic and subvolcanic rocks, as well as the feldspar-phyric metadacite dikes (Fig. 13; Table 4).

Spatial variation in mass changes during hydrothermal alteration

Simplified deposit maps, with a color scale on individual samples representing the intensity of mass changes for each major element affected by the hydrothermal system (Table 4), expressed in the same manner as above, are shown in Figure 14. The major shear zone along the northern margin of the massive sulfides separates the northern and southern structural domains at the deposit (Kampmann et al., 2016b; Fig. 2).

Mass changes for Si vary in space (Fig. 14A). Significant silicification occurs close to the massive sulfides in the QA rock and in the generally less altered volcanic rocks spatially separated from the deposit, including those in Falun SE. By contrast, losses of Si occur in the other altered silicate-rich rocks in both northern and southern structural domains. Fe is generally enriched (Fig. 14B). However, there is a slightly lower gain in this element in the northern relative to the southern domain. Lower values also characterize Falun SE. The enrichment in Fe can, at least partly, be ascribed to Febearing sulfide (i.e., pyrite, chalcopyrite, and pyrrhotite) mineralization in the altered silicate-rich rocks, especially in the QA rock. A gain in MgO of 20 to 60 wt % is conspicuous on all sides of the deposit but higher values are absent in Falun SE (Fig. 14C). The overall levels of Fe and Mg enrichment are lower in this area, the gain in Fe slightly exceeding that in Mg (Table 4; Fig. 14B, C). These results, which are consistent with a more peripheral location of Falun SE with respect to the central mineralization in the hydrothermal alteration system, indicate less chloritization.

The rocks immediately surrounding the mineralization at Falun are generally characterized by a loss of Ca, K, and Na (Fig. 14D-F, respectively) related to feldspar-destructive alteration. However, some samples of altered silicate-rich rock in the southern structural domain, located close to the folded marble-skarn unit to the south of the massive sulfides (Fig. 2), gained Ca and locally even K. The gain in Ca may have been related to a metamorphic reaction between this unit and the marble or skarn. The marked enrichment in K in Falun SE (Fig. 14E) is coupled to the stronger sericitization in this area.

Compositionally similar altered silicate-rich rocks preserve evidence for intense chloritization, silicification, and the breakdown of feldspar rocks to the east, south, and west of the massive sulfides in the southern structural domain, which is interpreted to represent the proximal footwall to mineralization. The rocks in the northern structural domain dominated by BQC schist (Fig. 2) also experienced strong chloritization, as well as feldspar breakdown and sericitization. However, since they are tectonically decoupled from the southern domain, it is difficult to correlate them with altered silicate-rich rocks south of the shear zone. They are interpreted to represent a more peripheral part of the intense footwall alteration envelope.

The current structural model for the Falun deposit, based on structural geometries and the distribution of altered silicaterich rocks at the deposit, involves one or several megascopic, steeply plunging sheath folds (Kampmann et al., 2016b). This model implies that the massive sulfides form the central part of these folds, surrounded by stratigraphic footwall rocks (Kampmann et al., 2016b). The spatial patterns of mass change during intense alteration discussed above, together with the presence of the same volcanic precursor rocks on all sides (Fig. 9F), support this structural model for the Falun deposit, which implies the absence of stratigraphic hangingwall rocks in the direct vicinity of the deposit (Kampmann et al., 2016b). Since the footwall rocks at Falun are associated with disseminated to semimassive Cu-Au mineralization, the identification of the spatial distribution of such rocks is crucial for near-mine exploration.

Mineral chemical data

In the same manner as for the lithogeochemical data, the EMP mineral chemistry data are presented with their median and MAD values (Table 5). All mineral chemistry data (wt %) are shown in the electronic supplementary material (Table A2).

The mineralogy of the altered silicate-rich rocks (i.e., quartz, biotite, cordierite, almandine garnet, anthophyllite, and minor andalusite, chlorite, and talc) is consistent with metamorphism of the hydrothermally altered rocks under amphibolite- and, at a later stage, retrogressive, greenschist-facies conditions (e.g., Gifkins et al., 2005). The mineral chemical data yield distinctly more Mg-rich mineral phases for BQC and TC schists, and Fe-rich equivalents for QA rock, supporting the lithogeochemical separation of predominantly Mg-enriched BQC and TC schists from Fe-enriched QA rock (Table 5; Fig. 15A).

The mass-change calculations are consistent with chloritization, sericitization, and silicification in the QA rock, as the predominant alteration styles in the hydrothermally altered zone at the Falun deposit. The following metamorphic reactions may take place during amphibolite-facies metamorphism of felsic volcanic rocks affected by such hydrothermal alteration (Barrett and MacLean, 1994; Barrett et al., 2005) and are consequently relevant for Falun (Lasskogen, 2010):

  • R1: Chlorite + quartz = cordierite + anthophyllite + water (QA rock and BQC schist)

  • R2: Chlorite + sericite + quartz = biotite + cordierite + anthophyllite + water (BQC and BC schists, and QA rock)

  • R3: Fe chlorite + sericite + quartz = biotite ± almandine garnet ± staurolite ± gedrite + water (QA rock and BA schist)

  • R4: Sericite + chlorite + quartz = cordierite + biotite + andalusite + water (QA rock)

To further test the hypothesis of sericite and chlorite as the predominant primary alteration minerals, the compositional range of chlorite at unmetamorphosed VMS deposits (Barrett et al., 2005) has been used for reference. The following analysis does not address the retrogressive mineral reactions that produced the chlorite and talc observed at Falun.

It is inferred that reactions R1 and R2 were possible in both the BQC schist and QA rock, since the respective tie lines for cordierite and anthophyllite pass through the assumed chlorite protolith field and lie close to the measured composition of biotite in these two rocks (Fig. 15A). The QA rock contains less biotite, suggesting a dominance of reaction R1 in this altered silicate-rich rock type. The positions of the tie lines for R1 and R2 in the QA rock indicate that the chlorite in the protolith was distinctly more Fe rich compared to the chlorite affected by the same reactions in the BQC schist.

Almandine garnet is common in some QA rocks, as well as in the biotite-almandine-(anthophyllite) schist. Based again on the position of the tie lines (Fig. 15A), Fe-rich chlorite, sericite, and quartz could have produced the observed QA and BA mineral associations according to reaction R3. Furthermore, chlorite in the protolith rocks to the QA rock and BA schist was more Fe rich in contrast to the protolith chlorite in the BQC, BC, and TC schists. This is in agreement with the mass-change results for these rocks, showing that the QA rock gained Fe to a larger degree than Mg (Table 4).

Andalusite is an accessory mineral phase, mainly observed in QA rock. Andalusite formed instead of anthophyllite, if significant sericite was present for reaction R4 to take place. The tie line for this reaction linking andalusite (not measured), biotite and cordierite (Fig. 15A) does not enter the chlorite field for unmetamorphosed VMS deposits, suggesting a slightly higher Al content in the initial chlorite composition than that used here.

Garnet at the Falun deposit has a similar visual character in the QA rock and BA schist. A mineral chemical profile in a garnet porphyroblast from a BA schist was generated to test for possible compositional zoning. In general, high FeO and MnO mass fractions, and low contents of CaO and MgO are present (Fig. 15B). The compositional zoning of the garnet (Fig. 15B) resembles typical metamorphic garnet growing along a clockwise P-T path, indicated by a decrease in CaO toward the rim, and prograde growth, indicated by increasing MgO and FeO, and decreasing MnO toward the rim (e.g., Tracy and Robertson, 1976). Since MnO mass changes in the altered silicate-rich rocks were generally low (Table 4), it is suggested that crystallization of garnet was triggered by the high Fe content in the QA rock and BA schist. In BQC schist, MnO was incorporated into anthophyllite (1.21 wt % in median values; Table 5), due to the absence of almandine garnet in this rock type.

BaO values in biotite from QA rock and BA schist are moderately elevated (0.29 and 0.62 wt % in median values, respectively; Table 5). In addition, F values of 0.75 wt % (median value; Table 5), with a few samples exceeding 1 wt % F (Table A2), have been recorded in biotite from QA rock. Even though these mass fractions are elevated in comparison with biotite from other altered silicate-rich rock types at Falun (Table 5), they lie within the normal ranges for biotite in metamorphic rocks (≤ 1 wt % BaO, ≤4 wt % F; Deer et al., 2012). Nonetheless, Ba-enriched micas are well known from several metamorphosed ore deposits worldwide, but the origin of Ba is a matter of debate (Majka et al., 2015).

Gahnite and staurolite have been identified as indicator minerals for ,metamorphosed massive sulfide deposits elsewhere (Sandhaus and Craig, 1986; Spry and Scott, 1986a, b; Spry et al., 2000). Gahnite is an accessory mineral in the altered rocks at Falun, preferably occurring in QA rock in close vicinity to the massive sulfides. Spinel at Falun contains 50 to 70 mol % ZnAl2O4 (gahnite), 16 to 50 mol % FeAl2O4 (hercynite), and 0 to 24 mol % MgAl2O4 (spinel) components (Wolter and Seifert, 1984). The dominance of the gahnite component (Spry and Scott, 1986a), as well as the occurrence in altered rocks in close vicinity to massive sulfides, are in agreement with results from metamorphosed VMS systems (e.g., Sandhaus and Craig, 1986).

If present, staurolite occurs in rocks together with cordierite (Wolter and Seifert, 1984). Staurolite from Falun is relatively poor in Mg (XFe = 0.87) and has a low ZnO (1.3 wt %) content (Wolter and Seifert, 1984). This is in contrast to higher average values for ZnO in staurolite at other metamorphosed massive sulfide deposits (5–9 wt %; Sandhaus and Craig, 1986; Spry and Scott, 1986b; Spry et al., 2000), formation of staurolite by desulfidation of sphalerite being suggested (Spry and Scott, 1986b). The low ZnO values in staurolite may reflect the low degree of mineralization in more peripheral cordierite-rich altered rocks at Falun (Fig. 2). The usefulness of gahnite and staurolite as indicator minerals for massive sulfide systems at Falun is questionable because of the general rarity of these minerals.

Premetamorphic character of the mineralization and hydrothermal alteration zoning

No evidence for exhalative sulfide deposition has been observed at Falun. Instead, textural evidence of premetamorphic replacement of carbonates (Fig. 5F) and volcanic rocks (relic quartz fragments; Fig. 5D) in the pyritic massive sulfides indicate subseafloor replacement for the massive sulfide mineralization, according to the criteria of Doyle and Allen (2003). Drill core mapping in this study and underground mapping by Geijer (1917) show that carbonate fragments are restricted to an approximately 10-m-wide margin in a girdle around the central massive sulfide mineralization (Fig. 2). Where such mineralization borders Cu-Au stockwork mineralization in the footwall, the transition is sharp (Geijer, 1917). Bearing in mind the spatial distribution of lithologies (Fig. 2) and the sheath fold model with tubular structure (Kampmann et al., 2016b), the carbonate distribution can be explained as relics that originally belonged to a single, laterally continuous unit in the stratigraphically lower part of the massive sulfide mineralization. However, the predominance of silicate over carbonate fragments in the massive sulfides, particularly in the central part of the deposit, suggests that most sulfides are hosted by volcaniclastic rocks deposited stratigraphically above this limestone unit.

The chemical and mineralogical variation in the hydrothermally altered volcanic rocks is interpreted to reflect different parts of a zoned footwall alteration system. In this context, the QA protolith represents a siliceous and intensely chloritized core of the hydrothermal system, as indicated by gains in Si, Fe, and Mg and the occurrence of disseminated Cu-Au mineralization, forming a stockwork zone in the central footwall (Table 4; Fig. 14). Iron-rich chlorite was dominant in the QA protolith, whereas the adjacent BQC protolith was characterized by Mg-rich chlorite. Both the spatial patterns of alteration and the premetamorphic chlorite composition represent empirical vectors toward mineralization and are therefore relevant for exploration.

It is suggested that the protolith to the BQC schist represents an intermediate alteration product between entirely chloritized QA protolith and more distal, sericitized, and silicified volcanic rocks. Lower K loss in the BQC schist compared to the QA rock (Table 4) implies that chloritization in the BQC protolith was less intense, the rocks containing sericite that could react with chlorite to form abundant biotite during metamorphism. Support for this alteration scheme is gained from the alteration box plot in which the trend from least altered volcanic rocks to sericitization is accompanied by the BQC trend from sericitization to chloritization (Fig. 7A, B). The more peripheral nature of the BQC protolith, with regard to the proximal alteration envelope, is also supported by its spatial distribution (Fig. 2). In the northern part of the deposit, this rock type is tectonically decoupled by a shear zone, and in the southern part only represented by a single, isolated body inside the QA rock.

The altered silicate-rich rocks in Falun SE are probably located along the gradational boundary to the central zone dominated by BQC schist and QA rock (i.e., the zone of metamorphosed strong chloritization). This is corroborated by the interfingering lenses of volcanic rocks and altered silicate-rich rocks, as well as the predominant sericitization and only moderate chloritization in this area (Fig. 14; Table 4). Mineralogy and alteration styles in Falun SE are similar to those in the BQC schist, with the exception that sericite has been overprinted to an even lesser degree by chloritization.

Composition and evolution of the hydrothermal fluid

A seawater source for the hydrothermal fluid is suggested by the significant Mg gain in the silicate-rich altered rocks (Baker and DeGroot, 1983; Allen et al., 1996; Franklin et al., 2005). Based on the dominance of ferrous iron minerals in the silicate-rich altered rocks (Table 5; Fig. 15) and the presence of Au mineralization and cassiterite together with Cu in the QA rock, it is inferred that the hydrothermal fluid was reducing (e.g., Cooke et al., 2000). The major role of ferrous iron and abundance of reduced sulfur in the fluid is further indicated by the pyritic character of the mineralization, the absence of hematite in mineralized and altered rocks, and the absence of an obvious sulfur source or reductant at the site of sulfide deposition. In such a fluid, sulfur may be transported, together with metals, mainly as H2S; SO42 plays a minor role (e.g., Cooke et al., 2000). Chlorite complexes may have been important for metal transport (see below). The possible presence of fluorine complexes is indicated by slightly elevated F concentrations in biotite (>1 wt % in a few samples; Tables 5, A2). However, this remains speculative since F content in the minerals at Falun is not unusually high (Deer et al., 2012).

In a reducing fluid, efficient metal transport requires low pH (Hannington et al., 1995, 1999; Cooke et al., 2000; Franklin et al., 2005), consistent with the feldspar-destructive alteration at Falun that is indicative of an acidic to moderately acidic fluid (pH ≤4). Efficient transport of Cu, which is one of the main commodities at Falun, requires high temperatures (300°–400°C) at acidic and reducing conditions (e.g., Large, 1992; Seyfried and Ding, 1995; Seyfried et al., 1999). The positive Eu anomaly in dolomite or calcite marble (Fig. 11I) also suggests involvement of a hot (>250°C), reducing, and acidic fluid that had liberated Eu2+ during alteration of feldspars (Sverjensky, 1984).

Even though more proximal and pyrite-bearing QA rocks have higher whole-rock S content than more peripheral BQC schist (Table 2), QA biotite, cordierite and anthophyllite (i.e., metamorphic products of chlorite) have constantly lower Mg/(Mg + Fe) than the BQC rocks (Table 5). Experimental work by Bryndzia and Scott (1987) suggests that Mg/(Mg + Fe) in chlorite would increase with increasing fS2, since Fe would preferentially enter pyrite. If chlorite Mg/(Mg + Fe) was simply a function of fS2, an Mg/(Mg + Fe) trend opposite to that observed at Falun would be expected from the center to the periphery of the hydrothermally altered zone. This lateral trend is instead easier to understand if formation of pyrite and Fe-rich chlorite in the core of the alteration system (QA protolith) caused an increase of Mg/(Mg + Fe) in the fluid from the center to the margins of the alteration system. In addition, the inferred temporal alteration sequence from Mgrich alteration (BQC) to more Fe-rich alteration (QA) implies that successive alteration caused an increase in sulfur content and a decrease of Mg/(Mg + Fe) in the proximal altered rocks with time.

Mode and mechanisms of mineralization

A conceptual model for the hydrothermal alteration and mineralization system at Falun is presented in Figure 16A. In order to illustrate the effects of later metamorphism and deformation of the system, including sheath folding along steep axes (Kampmann et al., 2016b), a north-south profile (Fig. 2), showing the current geometric relationships, is shown in Figure 16B. Bearing in mind that the mineralization at Falun is inferred to have formed from a hot, reducing and acidic hydrothermal fluid carrying metals and sulfur together, it is conceivable that upflow of hydrothermal fluid occurred along synvolcanic faults (Fig. 16A). For this reason, it is assumed in the conceptual model that the shear zones in the TC schist (Fig. 2) represent reactivated synvolcanic faults (Fig. 16B).

Under hot and acidic conditions, Au (in chloride complexation) and Cu solubility is mainly temperature dependent (Hannington et al., 1995, 1999; Franklin et al., 2005; Fig. 16C). For this reason, it is inferred that the precipitation of Cu and Au in the stockwork was mainly driven by cooling of the hydrothermal fluid during upflow (Point A, Fig. 16A, C). The porosity and permeability of the footwall pumice breccia at this stage of hydrothermal flow is unknown. Although an initial high porosity would have promoted fluid dispersion rather than focusing, the effects of burial compaction, diagenetic alteration, and earlier hydrothermal alteration would have decreased permeability. For example, the combined pattern in Figure 7A-B suggests that pervasive sericitization took place during the build-up phase of the hydrothermal system, followed by intense chloritization related to base metal deposition (Large et al., 2001).

As the fluid came into contact and reacted with carbonate stratigraphically above the stockwork mineralization, a rapid increase in fluid pH occurred. This would have facilitated both dissolution of limestone and pronounced decrease in Zn, Pb, and Cu solubility, whereby neutralization superseded cooling as the dominant precipitation mechanism (Seward and Barnes, 1997; Allen, 2010). Carbonate dissolution may have promoted downwelling of cooler seawater, which upon mixing with the hydrothermal fluid may have contributed further to sulfide deposition (Point B, Fig. 16A, B).

Based on the indication that the majority of the massive sulfide mineralization replaced volcaniclastic material above the former limestone, it is inferred that the hydrothermal fluid completely penetrated the limestone and that massive Zn-Pb- Cu-rich sulfide precipitation continued in overlying volcaniclastic sediments (Point C, Fig. 16A, B). If unconsolidated pumice breccia was the dominant host unit at this higher stratigraphic level, the massive nature of the mineralization in this zone could be explained by the high porosity (ca. 90 vol %) and seawater saturation in such volcaniclastic deposits prior to diagenesis (Doyle and Allen, 2003). Massive sulfides could then form if the rapid decrease in solubility of metals in the fluid due to mixing with cool seawater and a decrease in pH (Fig. 16C) outweighed the dispersing effect of a high porosity. Due to erosion of the stratigraphically higher portions of the massive sulfides and the hanging wall, this hypothesis cannot be verified. In addition to primary porosity in pumice deposits and carbonate dissolution, space for massive sulfide formation could also have been generated by active extensional faulting and fracturing, now represented by parts of the mineralization that lack relic rock fragments (Fig. 5E).

Besides cooling and neutralization, boiling has to be considered as another possible precipitation mechanism for sulfides at Falun. The facies associations at Falun are similar to those described by Allen et al. (1996) for other areas in the Bergslagen ore district, for which the paleoenvironment at the time of volcanism was inferred to be mainly shallow marine (≤ 200 m). A hydrothermal fluid with the properties deduced above would inevitably boil at water depths of less than 1,000 m (Butterfield et al., 1990; Hannington et al., 1999; Monecke et al., 2014). Boiling and associated phase separation is generally an efficient process for precipitation of metals in the subseafloor environment. This would lead to the precipitation of the entire metal budget of the fluid in vertically extensive (~1,500 m) zones of disseminated mineralization (Drummond and Ohmoto, 1985; Hannington et al., 1999). However, it is likely from the spatial affinity of massive sulfides to former carbonate rock that neutralization was a more efficient mechanism, and that the hydrothermal fluid retained sufficient metals to form massive sulfides up to the point where it interacted with the carbonate unit.

These considerations raise additional doubts that a significant seafloor component of the massive sulfide system ever existed, since it would have been difficult for the hydrothermal fluid to reach the seafloor in a shallow-marine environment without losing most, if not all, of its metal load. However, in a subseafloor setting, it is possible that the combined water pressure of the seawater column and seawater-saturated volcaniclastic sediments on top of the hydrothermal system were sufficient to prevent fluid boiling (Doyle and Allen, 2003). Processes of volcanotectonic subsidence and rapid burial could have aided in burying the limestone to a sufficient depth. This is in agreement with the results of Allen et al. (2003) from the SVALS deposit at Garpenberg. Here, the limestone, which later became the host to mineralization, underwent subsidence and burial by at least 500 m of pumice breccia during a major, caldera-forming volcanic eruption directly preceding the mineralizing event.

Source of sulfur and metals

A study of regionally alkali-altered volcanic rocks in the Bergslagen ore district showed that the regionally extensive zones of Na-altered volcanic rocks in the stratigraphically lower parts of the district are depleted in metals (Lagerblad and Gorbatchev, 1985). Such zones were included by Galley (1993) in the description of semiconformable alteration zones associated with VMS deposits. Derivation of metals from leaching of felsic volcanic rocks is consistent with results of a Pb isotope study (Sundblad, 1994), which demonstrated that ore Pb has an isotopic composition similar to that of the felsic volcanic rocks in the Bergslagen ore district. Values of δ34S in ore sulfides from Falun in the range –2.3 to –0.2%o were reported by Gavelin et al. (1960) and are compatible with volcanic rocks being the principal source of sulfur (Wagner et al., 2005). Lagerblad and Gorbatchev (1985) did not exclude a magmatic-hydrothermal contribution of metals from cooling intrusions to the ore-forming systems. However, the comagmatic relationship between volcanic and intrusive rocks indicated in this contribution implies that the isotopic data cannot fully discriminate between these two potential metal and sulfur sources.

In many hydrothermal systems, Au is regarded to have been derived from magmatic-hydrothermal fluids originating from intrusions that may also have functioned as heat engines for o driving hydrothermal convection of seawater (Poulsen and Hannington, 1996; Yang and Scott, 1996; Hannington et al., 1999; Dubé et al., 2007; Mercier-Langevin et al., 2011). In VMS settings, significant Au endowment is often associated with alteration reminiscent of subaerial epithermal systems (Dubé et al., 2007; Mercier-Langevin et al., 2011). At Falun, obvious equivalents to, for example, advanced argillic alteration have not been observed, although andalusite occurs as an accessory phase in strongly siliceous QA rock. A magmatichydrothermal contribution to fluids and metals, especially Au, is possible at the Falun deposit.

Origin of carbonate rocks

Elsewhere in the Bergslagen lithotectonic unit, carbonate rocks have been shown to represent metamorphosed stromatolitic limestone deposited in a shallow-marine setting during pauses in volcanic activity (Allen et al., 1996). Thick, laterally extensive carbonate units observed, for example, at the Garpenberg (Allen et al., 2003; Jansson and Allen, 2015) and Stollberg (Jansson et al., 2013) base metal sulfide deposits (Fig. 1A), are not present in the Falun inlier. An argument for a mainly biogenic origin of the carbonates at Falun rests in the relatively pure and massive appearance of the marble fragments in the massive sulfides (Franklin et al., 2005). However, a possible hydrothermal origin needs to be addressed since positive Eu anomalies occur in the marbles and skarns (Fig. 11I, J).

The PAAS-normalized (McLennan, 1989) REE data for carbonates from the Falun deposit are compared with data from the metamorphosed Stollberg deposit, Bergslagen ore district (Jansson et al., 2013), and with carbonate and sulfide mounds in active hydrothermal vent systems at Lake Tanganyika, East Africa (Barrat et al., 2000), and the TAG vent field, Mid-Atlantic Ridge (German et al., 1993), respectively. A hydrothermal origin of carbonate rocks was also discussed for the Stollberg deposit (Jansson et al., 2013; Fig. 17). All of these sediments show positive Eu anomalies, similar to the Falun data (Fig. 17).

A hydrothermal origin for the carbonates at Falun is not obvious from the mass-change calculations, which suggest that most rocks in the alteration system, including parts proximal to carbonate relics, were depleted in Ca. Furthermore, the bulk of the textural and chemical evidence favors replacement of carbonates rather than their formation in the alteration system. A hydrothermal origin would thus necessitate that carbonates were deposited early during the hydrothermal alteration, but were subsequently replaced when the hydrothermal fluids had attained higher temperatures and a lower pH. Additional data from C and O isotopes would help resolve the question around the biogenic or hydrothermal origin for these rocks.

Hydrothermal alteration associated with Zn-Pb-(Ag) skarn mineralization

Field work in the open pit suggests that tremolite skarn commonly occurs along the contact between dolomite marble fragments and altered silicate-rich rocks in the northwestern part of the deposit (Fig. 2). A reaction skarn model may account for the formation of this skarn. However, a reaction skarn model without aid of fluids is not able to explain thicker zones of pervasive carbonate replacement by massive lenses of diopside-hedenbergite skarn with associated Zn-Pb-(Ag) mineralization at Falun (Meinert et al., 2005). Instead, the latter is similar to alteration in metasomatic skarn systems (e.g., Meinert, 1992; Meinert et al., 2005). Assuming temperatures of ≤400°C, as suggested above for the hydrothermal fluid, these massive skarns could only have formed under exceptionally high fluid/rock ratios and efficient removal of CO2 from the hydrothermal system (Meinert et al., 2005; Jansson and Allen, 2015).

A similar conflict between VMS- and metasomatic skarnlike features were described at the marble and skarn-hosted Zn-Pb-Ag-(Cu) + magnetite deposit at Ryllshyttan (Jansson and Allen, 2015). Here, a hybrid model in which skarns formed at a transition from a subseafloor hydrothermal system to a contact metasomatic environment following emplacement of the earliest GDG intrusions was proposed (Jansson and Allen, 2015). This transition was induced by a combination of relatively continuous burial of the volcanic rocks and intrusion of the interbedded limestone by successively deeper intrusions, resulting in sequential alteration styles at increasing temperatures.

A model similar to that proposed for Ryllshyttan is consistent with the magmatic evolution at Falun and could also explain the formation of massive skarns at a late stage during the alteration process. However, specific intrusive bodies that can be linked to metasomatic skarn formation have not been identified at Falun.

Comparison to other SVALS deposits in the Bergslagen ore district

Mineralization at Falun involving at least some replacement of carbonate rock is similar to other SVALS deposits in Bergslagen, such as Garpenberg (Allen et al., 2003), Sala (Jansson, 2016), Ryllshyttan (Jansson and Allen, 2015), and Stollberg (Jansson et al., 2013). As at the Falun deposit (Kampmann et al., 2016a), a clear link to contemporaneous volcanism can be defined either stratigraphically (Allen et al., 1996, 2003, Jansson and Allen, 2015) or geochronologically (Jansson and Allen, 2011) at the Stollberg and Garpenberg deposits, where mineralization replaced limestone units deposited during pauses in the volcanism. Carbonate replacement is an important indicator for a chemical (pH) trap, but probably formed a subordinate component at Falun, in contrast to other SVALS deposits.

Crosscutting relationships at Garpenberg suggest that sulfide mineralization formed by subseafloor replacement of a stromatolitic limestone following a caldera-forming, explosive volcanic eruption (Allen et al., 2003). Geochronological data of Jansson and Allen (2011) indicate that sulfide deposition occurred in broad conjunction with, but mainly before, emplacement of subvolcanic intrusions, closely followed by the emplacement of GDG plutons, which is similar to the magmatic evolution at Falun (Kampmann et al., 2016a). At Stollberg, it was demonstrated that sulfide deposits formed in an intracaldera setting, and the replaced limestone unit directly overlies a ~500-m-thick pumice breccia, interpreted as a caldera fill deposit (Jansson et al., 2013). At Falun, pumice breccia forms a large portion of the footwall and probably also the ore-bearing strata. The presence of pumice deposits indicates a proximal volcanic source, although an intracaldera setting could not be substantiated using the current data set.

SVALS deposits that have exposed hydrothermally altered zones reveal highly asymmetric alteration aureoles, comprising more widespread zones of strong to intense alteration in the stratigraphic footwall, and weaker alteration in the stratigraphic hanging wall (Allen et al., 2003, Jansson et al., 2013). The direct hanging wall of the deposits may be intensely altered, but this alteration can be distinguished in style from the footwall alteration (e.g., Stollberg deposit; Jansson et al., 2013). The proximal parts of the alteration aureoles are generally characterized by schists rich in muscovite, biotite, and chlorite with porphyroblasts of garnet, cordierite, andalusite, sillimanite, or staurolite, depending on composition and metamorphic grade. These rocks are characterized by intense Na depletion and variable enrichment in K, Fe, Mg, and Si, and locally host stockwork-type Cu-Zn mineralization (Vivallo, 1984; Allen et al., 2003; Jansson et al., 2013). Some SVALS deposits have deeper zones of Na enrichment (e.g., Stollberg deposit. It has been argued that this reflects temperature zoning within a convective, modified seawater-dominated hydrothermal system (Ripa, 1988, 2012; Jansson et al., 2013).

Zones of Na enrichment could not be identified at the Falun deposit. Sodium, Ca, and K were generally depleted during hydrothermal alteration in the altered silicate-rich rocks, with some K enrichment due to sericitization in the more peripheral rocks of Falun SE, and also in the volcanic and subvolcanic rocks of the Falun inlier (Fig. 14; Table 4). Due to the absence of hanging-wall exposure at the deposit, it is not known if a similar asymmetry in alteration intensity exists as at Garpenberg and Stollberg.

The presence of Mn-rich garnet and anthophyllite, and minerals with essential Mn has been reported at the Stollberg deposit (Ripa, 2012; Jansson et al., 2013) and in the Garpenberg area (Allen et al., 2003; Jansson and Allen, 2015). Jansson (2016) noted the absence of significant Mn enrichment at the Sala Zn-Pb-Ag deposit and speculated that this may reflect incomplete exploration or preservation of the alteration system. Similarly, no clear zones of significant Mn enrichment can be defined in the sampled portions of the Falun deposit, despite elevated Mn content in individual garnet porphyroblasts (Fig. 15B).

Whereas a genetic model bearing similarities to VMS and skarn mineral formation as a consequence of regional metamorphism has been suggested for Stollberg (Ripa, 1988; Jansson et al., 2013) and Garpenberg (Vivallo, 1984), the SVALS deposits at Garpenberg, Ryllshyttan, and Sala have many features in common with metasomatic skarn deposits (Allen et al., 2003; Jansson and Allen, 2015; Jansson, 2016). These include: (1) intimate association between skarn minerals (e.g., pyroxene, garnet, amphibole) and sulfides that replaced former limestone; (2) broadly coeval intrusive magmatic activity (although direct genetic links to specific intrusions cannot generally be proven); and (3) local zones in which former limestone was completely replaced by clinopyroxene and garnet skarns. At the Falun deposit, the major portion of the skarns is not spatially associated with the massive sulfide mineralization (Fig. 2), and calc-silicate fragments are rare in the latter. An exception is a large carbonate fragment in the northwestern part of the massive sulfides, which has a selvage of tremolite skarn in the contact zone to the massive sulfides (Fig. 2). The disseminated Zn-Pb-(Ag) mineralization occurring in both diopside-hedenbergite and tremolite skarns is inferred to be linked to the skarn alteration. Overprinting relationships between barren diopside-hedenbergite and mineralized tremolite skarns, as reported at the Ryllshyttan deposit (Jansson and Allen, 2015), were not observed. A model of separate, later skarn formation involving burial and metasomatism is preferred for the Falun occurrence.

Falun differs from the SVALS deposits at Garpenberg, Ryllshyttan, and Stollberg mainly on the basis of the following three features: (1) high pyrite content in the massive sulfides, (2) absence of related Fe oxide deposits, and (3) dominant replacement of volcaniclastic sediments compared to carbonates. Possibly, (1) and (2) reflect hydrothermal alteration and mineralization from a more H2S-rich hydrothermal fluid. In such a scenario, it may be that most Fe was sequestered in sulfides and silicates, leaving insufficient Fe to form peripheral Fe oxide deposits.

Comparison with VMS deposits

The results presented in this paper motivate a comparison with a VMS model of ore formation (e.g., Franklin et al., 1981, 2005; Lydon, 1984, 1988; Galley et al., 2007) for the Falun base metal sulfide deposit. Such a model involves convection of modified seawater below the seafloor and enrichment of leached metals in a zone of concentrated upflow, commonly along synvolcanic faults, due to heating from subvolcanic intrusions. In most cases, sulfide precipitation was triggered by fluid cooling and mixing with neutral seawater. This occurred commonly, especially for Cu-rich systems, in deep-marine settings where the water column pressurized the fluid system and prevented boiling (Monecke et al., 2014). Seafloor-exhalative mounds and stratiform sheets, and Curich stockworks associated with chloritization in hydrothermally altered zones in the subseafloor realm are the dominant mineralization styles.

The specific arguments for the comparison with a classic VMS genetic model for the Falun deposit are as follows: (1) location of the deposit in a host unit of predominantly felsic volcanic rocks; (2) rapid evolution (≤ 10 m.y.) from felsic volcanism to hydrothermal ore formation, late- or postmineralization emplacement of feldspar-phyric dacite dikes and plutonism (Kampmann et al., 2016a); (3) magmatic vent- proximal location for the deposit indicated by abundant hypabyssal intrusions of feldspar-phyric dacite and mafic dikes; (4) zonation from Cu-Au mineralization (stockwork) to pyritic massive sulfides, identified after restoration (unfolding) of an inferred sheath fold structure (Kampmann et al., 2016b); (5) prevalence of hot, acidic, and reducing fluids responsible for hydrothermal alteration and ore formation (Seyfried et al., 1999; Cooke et al. 2000; Franklin et al., 2005); (6) style of hydrothermal alteration in the inferred footwall with intense sericitization, chloritization, and feldspar breakdown, and silicification in the central part (QA rock); (7) mineral associations in the altered silicate-rich rocks consistent with metamorphic equivalents to common mineral associations at VMS deposits; and (8) spatial association with a major mineralized shear zone, which potentially formed a conduit for metal-bearing fluids, i.e., as a synvolcanic fault. Furthermore, mass-change calculations on footwall rocks at Falun (variable Si; enrichment in Mg and Fe; depletion in Ca, K, and Na, and insignificant mass change in Mn and P) are similar to the results at other metamorphosed, Paleoproterozoic VMS deposits, including Ruostesuo, Vihanti-Pyhäsalmi district, central Finland (Roberts et al., 2003), and Kristineberg, Skellefte district, northern Sweden (Barrett et al., 2005).

In detail, the geologic features at Falun are most similar to the subcategory of VMS deposits showing subseafloor replacive styles after reactive hosts including pumiceous or carbonate rocks (Allen and Weihed, 2002; Doyle and Allen, 2003; Tornos et al., 2015). A hybrid model between carbonate replacement and sub- or seafloor VMS has been suggested for the Lewis Ponds massive sulfide deposit, New South Wales, Australia. However, at Lewis Ponds, these components form separate zones and not a continuous system as at Falun (Agnew et al., 2005).

Although typical for SVALS deposits in the Bergslagen ore district, VMS deposits with associated skarn components, as at Falun, are rare (Lydon, 1988; Franklin et al., 2005). The Montauban polymetallic massive sulfide deposit, Quebec, Canada, comprises calc-silicate rocks, including diopside- and tremolite-bearing assemblages, in close spatial association with massive sulfide mineralization (Bernier and MacLean, 1993). The altered carbonates were interpreted as metamorphosed lenses of marine carbonate within felsic volcanic rocks. However, evidence for premetamorphic carbonate replacement by massive sulfides was not described by Bernier and MacLean (1993).

The host-rock suite of mainly rhyolitic and dacitic volcanic and subvolcanic rocks to the metamorphosed Falun Zn-Pb-Cu- (Au-Ag) deposit is largely comagmatic with coeval feldspar- phyric metadacite dikes at the deposit and granitoid plutons in the area, suggested by common magmatic fractionation trends. All these suites have calc-alkaline affinities. Volcanic rocks were affected by strong chloritization, sericitization, and silicification in the stratigraphic footwall to the deposit, which resulted in an intense mineralogical modification in this zone to premetamorphic assemblages, including proximal Fe-rich or more peripheral Mg-rich chlorite, as well as sericite and quartz. During subsequent lower amphibolite-facies metamorphism, the hydrothermal mineralogy was transformed to mineral associations containing variable amounts of biotite, quartz, anthophyllite, cordierite or almandine, and accessory andalusite, gahnite, staurolite, and gedrite.

Intensely chloritized footwall rocks, mainly representing a single, rhyolitic precursor with similar mineralogy coupled to similar mass changes during alteration, envelop the mineralization on all sides. The northern side is tectonically decoupled from the southern side and the massive sulfides by a major shear zone. The general concentric geometric pattern is consistent with a previously suggested model involving sheath fold formation with the massive sulfides as the stratigraphically highest preserved unit in the center surrounded by stratigraphic footwall rocks in a tubular structure (Kampmann et al., 2016b). These results are also relevant for exploration, since footwall rocks at Falun host disseminated to semimassive stockwork Cu-Au mineralization.

Ore formation by circulation of modified seawater occurred in the subseafloor regime within a succession of porous pumice breccia and limestone. Synvolcanic faults, later modified to ductile shear zones, may have facilitated upflow of hot, acidic, reducing, and metal-bearing fluids. Following Cu-Au stockwork mineralization by fluid cooling, the fluids encountered limestone stratigraphically higher up in the system. This gave rise to neutralization of the metal-bearing fluids and a shift toward a pH-controlled precipitation of Zn-Pb-Cu-rich massive sulfides. Dissolution processes, primary porosity in the pumice breccia, and secondary porosity produced during synvolcanic faulting could have all contributed to creation of the space necessary for the formation of the massive sulfides.

The character of the host rock and hydrothermal alteration, the type and zonation of mineralization, and the inferred metamorphic reactions at Falun encourage a comparison with metamorphosed, pyritic VMS deposits. An inferred premetamorphic alteration zonation from distal sericite-altered and silicified volcanic rocks to intermediate sericite- and chloritealtered and proximal siliceous and intensely chloritized rocks is in agreement with this interpretation, and represents an alteration vector toward mineralization. The importance of carbonate replacement and the spatial association with subordinate skarn alteration are features that differ from a classic VMS model.

The Geological Survey of Sweden (SGU) and Luleå University of Technology provided the financial support for this study. Magnus Ripa (SGU) is thanked for providing the geologic mapping results from the Falun area, and Stefan Luth, Per Nysten, and Magnus Ripa (all at SGU) participated in the sampling of the rocks surrounding the Falun deposit in connection with this geologic mapping work. All are thanked for their general support during the field studies. Philipp Kampmann and Nathalie Pérez assisted in generating mass-change calculation spread sheets and figures, respectively. The Falu Gruva Foundation granted access to the open pit and is thanked for permission to collect samples. The SGU branch office in Malå, Sweden, as well as the exploration company Drake Resources, provided access to drill core and permission to sample the cores. Employees at Boliden AB in Garpenberg are acknowledged for technical help with sample preparation. This paper benefited greatly from reviews by Mark Hannington, Paul Duuring, Paul Spry, and Larry Meinert.

٭
A digital supplement containing electronic Tables A1 and A2 for this paper is available at http://economicgeology.org/ and at http://econgeol.geoscience world.org/.

Supplementary data