Age and Chemostratigraphy of the Finlayson Lake District, Yukon: Implications for Volcanogenic Massive Sulfide (VMS) Mineralization and Tectonics along the Western Laurentian Continental Margin

The Yukon-Tanana terrane represents the largest arc–back-arc terrane in the northern Cordillera and is a critical component for deciphering the early tectonomagmatic and metallogenic evolution of the peri-Laurentian realm [1, 2]. Rocks that comprise Yukon-Tanana terrane have geochemical, isotopic, and metallogenic links to the Laurentian continent (e.g., [3–6]) yet display magmatic, metamorphic, structural, and sedimentary characteristics that differentiated it from the neighboring platformal strata, thus defining an allochthon-parautochthon relationship [1, 7]. The Finlayson Lake district is the most stratigraphically intact segment of Yukon-Tanana terrane rocks in the northern Cordillera and has been the locus of research since the mid-1980s [4, 8]. There was increased interest in the area due to the discovery of numerous volcanogenic massive sulfide (VMS) occurrences in the mid-1990s [9–11], which has resulted in the discovery of >40 Mt of polymetallic ore, including the ~18.1 Mt Kudz Ze Kayah, ~1.5 Mt GP4F, ~10 Mt Kona, and 6.2 Mt Wolverine VMS deposits [12–14]. Coincident mapping, geochemical, tracer isotope, and U-Pb geochronological work defined the chronology and chemostratigraphy of the Finlayson Lake district and illustrated that it formed as an evolving arc–back-arc system between ca. 366 Ma and 345 Ma [15–17].

The geochronological work that underpinned much of the previous work was based on traditional, multigrain zircon fractions and air abrasion methods [18] prior to significant advancement in the field of U-Pb geochronology with the invention of chemical abrasion pretreatment techniques [19]; thus, issues remain with regard to the accuracy and resolution of various plutonic and volcanic events and the timing of VMS mineralization in the Finlayson Lake district (e.g., [20]). Advances in high-precision U-Pb geochronology have revolutionized our understanding of ore systems and have allowed refinement of the timing of volcanic and plutonic activity related to mineralization on relatively short time scales from 10 s to 100 s of thousands of years [21–25]. Refinement of the chemical abrasion pretreatment technique [19] and application of the EARTHTIME standard [26] to U-Pb zircon geochronology have increased precision and relative uncertainties to as low as ~0.1% for single crystals and ~0.02% for weighted means, compared to laser ablation methods that achieve closer to ~1% [27]. In VMS districts, this precision provides the ability to define mineralized horizons on a regional scale and assess the timing of past volcanic eruptions and plutonic activity and their relationships to VMS mineralization [28, 29]. In the Finlayson Lake district, these high-precision geochronology methods have defined the ages of the Kudz Ze Kayah, GP4F, and Wolverine VMS deposits [30]; however, the age relationship to local plutons with respect to the proposed heat sources to the VMS systems remains unclear.

Magmatism is critical in the formation of VMS deposits, with magmatic heat interpreted to be the thermal driver of VMS-related hydrothermal circulation in the upper crust [31–34]. While it remains uncertain if plutons and subvolcanic intrusions spatially associated with VMS camps provide enough heat to form large deposits [35, 36], especially in light of recent models of incremental pluton assembly in arc settings [37–40], it is clear that high-temperature felsic volcanism is commonly associated with VMS mineralization in continental-dominated terranes [17, 41–44]. This high-temperature magmatism is often expressed as high concentrations of high field strength elements (HFSE; e.g., Nb, Th, Y, and Zr) and rare earth elements (REE; [17, 41, 43, 45]) and has been suggested that this is due to basaltic underplating of continental crust that provides higher than normal geothermal gradients, induces greater dissolution of HFSE-REE minerals in the crust during melting, and by association has the potential for the genesis of large VMS deposits [17, 36, 41, 43]. To test this underplating-crustal melting hypothesis, petrological (i.e., MLA-SEM) and radiogenic tracer isotope geochemical data is critical for validating and quantifying the relative abundance of crust versus mantle components involved in VMS-related felsic magma genesis; however, this requires both detailed lithostratigraphic and chronostratigraphic controls and VMS-bearing and VMS-barren stratigraphy with similar continental basement, to properly assess magmatic controls on VMS genesis (e.g., [42, 46]).

In this contribution, we report new high-precision CA-ID-TIMS and LA-ICP-MS U-Pb zircon dates, whole-rock lithogeochemistry, MLA-SEM modal abundances, and Hf-Nd isotopes for the earliest arc–back-arc environments that formed immediately adjacent to the western Laurentian continental margin in the Late Devonian to Early Mississippian. Specifically, we report new results for back-arc rocks in the Big Campbell thrust sheet and arc rocks in the Cleaver Lake thrust sheets in the Finlayson Lake district of the Yukon-Tanana terrane. We present a revised interpretation of Yukon-Tanana terrane evolution that incorporates early and recent interpretations of the ultramafic rocks and their volcanic cover sequences in the district (e.g., [4, 8, 47–50]). In this new context, our results constrain the timing and petrogenesis of upper-crustal plutonism and volcanism in the back-arc rift and provide constraints on the timing, duration, and host rock composition of VMS mineralization at or below the seafloor. In the arc regions, these new data increase temporal resolution for the earliest-known arc magmatism in the district and further constrain the longevity of the arc during a period of active back-arc extension. These data significantly contribute to our understanding of the nature and timescales of magma-rich arc rifting and associated VMS mineralization along an established and attenuated continental margin setting along the ancient Cordilleran margin of Yukon, Alaska, and British Columbia.

The Finlayson Lake VMS district of southeastern Yukon is a fault-bounded block of the Yukon-Tanana and Slide Mountain terranes that formed off the western continental margin of Laurentia in the mid- to late-Paleozoic (Figure 1; [8, 16, 49]). Rocks that comprise Yukon-Tanana terrane arc and back-arc assemblages are variably deformed and metamorphosed volcanic, plutonic, and sedimentary rocks that locally retain primary geological and geochemical characteristics; these rocks were deposited or intruded above a pre- to Late Devonian basement [1, 5, 7, 16]. The Jules Creek transform fault juxtaposes the Yukon-Tanana terrane adjacent to ophiolitic rocks of the Slide Mountain terrane [16], which were then together thrust above North American platformal strata along the Inconnu thrust in the Late Jurassic to Early Cretaceous [51]. The current geographical configuration of the Yukon-Tanana terrane in central Yukon was attained in the Eocene following ~430 km of displacement along the Tintina strike-slip dextral fault system, which resulted in the offset sliver referred to as the Finlayson Lake district (Figures 1 and 2; [52]).

In the current interpretation of the structural geometry of the Yukon-Tanana terrane [16], Yukon-Tanana terrane rocks within the Finlayson Lake district are hosted in three distinct structural panels (from structurally deepest to shallowest): the Big Campbell, Money Creek, and Cleaver Lake thrust sheets (Figures 2 and 3; [16]). Upper Devonian and older metasedimentary rocks of the North River formation comprise the basement to both the Big Campbell and Money Creek thrust sheets, but the formation is not observed in the Cleaver Lake thrust sheet. The Big Campbell thrust sheet is bounded below by the post-Late Triassic Big Campbell thrust fault and above by the Money Creek thrust fault (Figures 2 and 4). Rocks in the Big Campbell thrust sheet comprise the basal Upper Devonian and older North River formation and the overlying Upper Devonian mafic and felsic metavolcanic and metasedimentary rocks of the Grass Lakes group, which includes the Fire Lake, Kudz Ze Kayah, and Wind Lake formations (Figure 2). Late Devonian gabbroic to dioritic intrusions of the North Lakes intrusion and granitoid intrusions of the Grass Lakes plutonic suite cut the Grass Lakes group and then are all unconformably overlain by Lower Mississippian metaclastic and mafic to felsic metavolcanic rocks of the Wolverine Lake group (Figures 3 and 4). Metasedimentary rocks of the Money Creek formation and metabasalt of the Campbell Range formation, both originally thought to be Lower Permian, are interpreted to unconformably overlie the Wolverine Lake group; we revisit and substantially revise this interpretation below. The Grass Lakes and Wolverine Lake groups contain VMS mineralization felsic volcano-sedimentary stratigraphy that is interpreted to have formed in an evolving continental arc to back-arc basin tectonic setting that represents the earliest stages of mid-Paleozoic rifting on the western peri-Laurentian margin [7, 16, 30, 47, 48, 53].

In the hanging wall of the Money Creek thrust fault, rocks comprising the Money Creek thrust sheet include the Upper Devonian North River formation that is overlain by Upper Devonian to Lower Mississippian felsic to intermediate metavolcanic and metasedimentary rocks (Waters Creek and Tuchitua River formations). These rock units are intruded by granitic rocks of the Early Mississippian Simpson Range plutonic suite, then capped by Mississippian to Lower Permian limestone, mafic metavolcanic, and metaclastic rocks (Figures 2 and 3; [3, 4, 16]). The rocks in the Money Creek thrust sheet are interpreted to have formed in a continental arc (±local back-arc basin) tectonic setting to the southwest of the Big Campbell thrust sheet and are presently not associated with any significant VMS mineralization [3, 7, 16]; however, several mineral occurrences are present [54]. Rocks from the Money Creek thrust sheet were not investigated in this study and will not be further discussed.

The Cleaver Lake thrust sheet structurally overlies the Money Creek thrust sheet and contains relatively undeformed and unmetamorphosed Late Devonian mafic and felsic volcanic rocks (Cleaver Lake formation) that overlie mafic and ultramafic rocks; these rock units are subsequently intruded by Early Mississippian granitoids of the Simpson Range plutonic suite [16, 49, 55]. The Cleaver Lake thrust sheet was thrust above the Money Creek thrust sheet along the Cleaver Lake thrust fault after the Early Permian [16]. Mafic-ultramafic rocks were initially interpreted to have been Permian intrusions that were correlative to the Slide Mountain terrane, thereby constraining the age of thrusting to Permian [16]; however, in this contribution, we revise this interpretation as discussed below. The rocks in the Cleaver Lake thrust sheet are interpreted to have been generated in an intraoceanic arc and subduction complex to the southwest of the Money Creek and Big Campbell thrust sheets and are also not associated with any known VMS mineralization [3, 6, 7, 16, 45].

Slide Mountain terrane rocks are juxtaposed against Yukon-Tanana terrane rocks along the Jules Creek fault (Figure 2; [16]). North and east of the Jules Creek fault, these rocks comprise Early Mississippian to Lower Permian metasedimentary and metavolcanic rocks of the Fortin Creek Group, which are unconformably overlain by Lower Permian mafic metavolcanic rocks and metasedimentary rocks of the Campbell Range Formation and Lower to Middle Permian sedimentary rocks of the Gatehouse formation (Figure 2). The Slide Mountain terrane contains the Lower Permian mafic-type Ice VMS deposit and has been interpreted to form in a back-arc basin to mid-ocean ridge tectonic setting [56–58].

The Fire Lake formation has traditionally been interpreted to be the stratigraphically lowest volcanic unit overlying the metaclastic rocks of the North River formation [16, 51]. The original definition of the formation comprised mafic volcanic-volcaniclastic rocks with geochemical affinities equivalent to enriched midocean ridge basalt (E-MORB), back-arc basin basalt (BABB), alkalic ocean island basalt (OIB), and low-Ti island arc tholeiite (IAT) to boninite affinities [7, 47, 48]; these mafic volcanic rocks overlie mafic-ultramafic bodies that were originally inferred to be comagmatic intrusions, an interpretation that is revisited herein. Variable Nd isotopes (⁠$εNd350Ma=–0.3$ to +8.5) and trace element signatures suggest they represent varying mixtures of back-arc mantle and arc mantle wedge with or without slab and/or continental crustal contributions that were generated during the onset of back-arc–related rifting along the Laurentian margin [47, 48]. The ~10 Mt Kona Cu-Co-Au VMS deposit is hosted by mafic volcaniclastic rocks (primarily boninitic) and fine-grained turbiditic metasedimentary rocks that are interpreted to have formed in a subaqueous setting [20, 59]. The age of this unit was originally thought to have been pinned by two Late Devonian ²⁰⁷Pb/²⁰⁶Pb dates (⁠$366.3±10.2$ Ma and $365.0±1.3$ Ma; [16]) for the North Lakes intrusion, which includes metadiorite-gabbroic rocks originally inferred to be comagmatic with Fire Lake volcanic rocks; we revisit and revise this interpretation below.

The Kudz Ze Kayah formation, originally thought to stratigraphically overlie the Fire Lake formation (see below), contains ~400-1300 m of felsic volcaniclastic rocks and subordinate flows, felsic and mafic subvolcanic intrusive rocks, and mudstones deposited in a rapidly forming, subaqueous back-arc basin [30]. Felsic lithofacies have evolved Nd isotopic signatures (⁠$εNd350Ma=–8$ to –9) but have HFSE and REE enrichment characteristic of A-type, back-arc magmatic rocks; they are interpreted to have formed from melting of Proterozoic to Archean continental crust [6, 45]. The felsic-siliciclastic ~1.5 Mt GP4F and bimodal-felsic ~18.1 Mt Kudz Ze Kayah Zn-Pb-Cu-Ag-Au deposits are hosted in felsic volcaniclastic rocks approximately 4 km laterally and ~500-600 m stratigraphically apart (Figure 4; [13, 30, 60]). The age of felsic volcanism in the Kudz Ze Kayah formation has recently been tightly constrained to ca. 363.3 to 362.4 Ma with four CA-ID-TIMS dates [30]. These new dates define the age of VMS mineralization in the GP4F deposit to $363.254±0.098$ Ma, whereas the Kudz Ze Kayah deposit is constrained by two overlapping dates in the Krakatoa and ABM zones (⁠$362.847±0.099$ Ma and $362.82±0.12$ Ma, respectively).

The Wind Lake formation overlies the Kudz Ze Kayah formation and consists of carbonaceous argillite; lesser mafic volcaniclastic, volcanic, and intrusive rocks; and minor felsic volcaniclastic rocks and quartzite. The total thickness of the Wind Lake formation is estimated at ~4500 m; mafic volcanic rocks are only present in the lowest 500 m and are otherwise dominated by carbonaceous argillite (Figure 4; [30, 60]). The mafic rocks have alkalic, OIB-type geochemical signatures with near-chondritic Nd isotope signatures (⁠$εNd350Ma=–2.8$ to +1.1; [61]), suggesting the magmas underwent varying degrees of crustal contamination following extraction from lithospheric or asthenospheric sources. The Wind Lake formation thus represents the cessation of felsic-dominant volcanism in the Grass Lakes group (Figure 4; [61]).

The Grass Lakes plutonic suite comprises a batholith-like intrusive complex composed of variably metamorphosed and deformed, fine to medium-grained peraluminous granodiorite to monzogranite (Figures 2, and 4; [16]). Minor intrusions (dikes and sills) of $feldspar±quartz$ porphyritic rocks cut the entire Grass Lakes group stratigraphy (Figure 4; [60]). The geochemical characteristics of this rock suite are similar to felsic, A-type, back-arc–related volcanic rocks of the Kudz Ze Kayah formation, with similarly evolved Nd isotope compositions (⁠$εNd350Ma=–9.5$⁠; [6]) and inherited Proterozoic to Archean zircon grains [4, 6, 16, 45]. The granitic rocks of the Grass Lakes plutonic suite range include dates of $362.2±3.3$⁠, $359.9±0.9$⁠, and $357.3±2.8$ Ma [16]. Geochemical and geochronological constraints have led previous workers to suggest a comagmatic relationship between silicic plutonism of the Grass Lakes plutonic suite and the formation of felsic volcanic rocks and associated VMS mineralization in the Kudz Ze Kayah formation [6, 45].

Rocks of the Wolverine Lake group unconformably overly the Grass Lakes group following a period of deformation during the Upper Devonian to Lower Mississippian, the kinematic nature of which is poorly known (ca. 357.5 Ma; [16, 62]). Rocks that comprise the Wolverine Lake group include undifferentiated basal, footwall, and hanging wall units [16, 30, 63, 64]. The basal unit, constrained by ca. 357.5 Ma maximum depositional age [16], contains quartzofeldspathic-pebble conglomerate, grit, sandstone, and carbonaceous phyllite. Footwall rocks comprise ~1200-1700 m carbonaceous argillite, felsic volcaniclastic rocks, and high-level quartz-feldspar porphyritic (QFP) and feldspar porphyritic (FP) intrusive rocks (Figure 4). Felsic crystal tuffs in the lower footwall have been previously dated at $356.371±0.091$ Ma [30] and $356.2±0.9$ Ma [4, 16], and two periods of magmatism are associated with VMS mineralization at (1) ca. 356.1 Ma: pre-VMS QFP intrusions (Sable zone) and (2) ca. 355.2 to 355.0 Ma: syn- to post-VMS FP intrusions and crystal tuffs (Wolverine/Lynx and Fisher zones; [30]). An additional QFP from the Puck zone gave an older date of $356.9±0.5$ Ma [17]. Geochemical and isotopic characteristics of the footwall felsic rocks (⁠$εNd350Ma=–8.2$ to –7.8; [6]) are similar to A-type, back-arc affinities of the Kudz Ze Kayah formation and Grass Lakes plutonic suite, with the exception of the porphyritic intrusive rocks where FP rocks are HFSE-REE-rich compared to QFP and volcaniclastic rocks [17, 45]. Hanging wall rocks primarily contain carbonate and silica-pyrite exhalite, iron formation, and carbonaceous argillites that are overlain by fine-grained resedimented rhyolitic tuff/siltstone and rhyolitic breccia and massive basalt flows and minor mafic volcaniclastic rocks with juvenile $εNd350Ma=+6.9$ N-MORB, enriched- (E-) MORB, and BABB geochemical affinities [6, 16, 30, 53, 64, 65]. The basalts in the Wolverine Lake group hanging wall were initially mapped as being part of the Campbell Range formation of the Slide Mountain terrane, an interpretation which is reconsidered in a subsequent section.

In the core region of the Finlayson Lake district, the Cleaver Lake thrust sheet is a klippe that contains two primary, relatively undeformed geological units: (1) the Upper Devonian Cleaver Lake formation and (2) Early Mississippian Simpson Range plutonic suite [16, 55]. The Cleaver Lake formation consists of both calc-alkaline and island arc tholeiitic basalt and rhyolite that are intruded by quartz-rich porphyritic rocks; the latter intrusions show local evidence for magma mingling with basaltic dikes and are interpreted to be the volcanic feeders to the extrusive rocks [55]. Geochemically, these rocks have arc-like signatures, and Nd isotopic signatures (⁠$εNd350Ma=–4.8$ to +0.1; [6]) indicate this arc was built above a basement composed of both oceanic and continental rocks [3, 6, 45, 48]. Rocks in the Cleaver Lake formation have crystallization ages of $360.5±1.9$ Ma and $356.1±0.9$ Ma for felsic and mafic intrusive rocks, respectively, and indicate that the primary volcanic succession is likely coeval in age to the Grass Lakes group [4, 7, 16]. The Simpson Range plutonic suite contains hornblende- and biotite-bearing monzodiorite, granodiorite, and quartz diorite, all with evolved geochemical signatures of arc affinity (⁠$εNd350Ma=–12.9$ to –7.4; [3, 6]); these granitoids exhibit cross-cutting relationships with rocks of the Cleaver Lake formation [3, 55]. U-Pb zircon dates for three granitoids in the Simpson Range plutonic suite include $354.9±1.8$⁠, $348.4±0.8$⁠, and $345.9±1.2$ Ma [3, 4, 16] and define a clearly younger age than the Cleaver Lake formation.

In the southeastern Finlayson Lake district, the Cleaver Lake thrust sheet also includes the Klatsa metamorphic complex. The Klatsa metamorphic complex comprises bodies of serpentinized mafic-ultramafic rocks structurally imbricated with eclogite-facies sedimentary rocks and basalt and represents an Early Mississippian subduction complex [15, 66].

In this study, we revise the stratigraphic and structural framework of the Finlayson Lake district presented above in light of our new results and the adoption of the reinterpretation of mafic and ultramafic bodies in the region presented by van Staal et al. [50], which follows early work by Tempelman-Kluit [49], Mortensen and Jilson [8], and Mortensen [4]. Geological observations are presented here as a basis for reporting geochronological, geochemical, and isotopic data below and described further in the discussion.

Traditionally, the Fire Lake formation has been considered to be in stratigraphic succession between the North River and Kudz Ze Kayah formations, which contained mafic-ultramafic bodies that were interpreted as coeval intrusions. However, we have adopted the reinterpretation by van Staal et al. [50] that follows others [4, 8, 49], where the ultramafic-mafic slabs represent allochthonous slices of mantle and lower crust as opposed to intrusions [11, 16]. This reinterpretation requires that the base of the mafic-ultramafic bodies be a thrust fault, and the relationship of the Fire Lake formation to the North River and Kudz Ze Kayah formations therefore requires reevaluation. The Fire Lake formation and underlying ultramafic-mafic rocks define a new thrust sheet that overlies the North River formation and also possibly the Kudz Ze Kayah formation if the North Klippen is correlated with the Fire Lake thrust sheet (Figure 2). This reinterpretation indicates that (1) the Fire Lake formation no longer stratigraphically overlies the North River formation but instead is structurally juxtaposed against it (Figure 5), and (2) the mafic rocks of non-arc affinity (BABB, E-MORB, and OIB) between the Kudz Ze Kayah and North River formations, initially correlated with the Fire Lake formation [48, 67–69], are no longer in the same thrust sheet and thus not directly correlatable to the Fire Lake formation. The reassignment of these rocks to the Kudz Ze Kayah formation makes the formation a bimodal volcanic succession that directly overlies the North River formation, which includes BABB and E-MORB-affinity mafic rocks with juvenile isotopic signatures (⁠$εNd350Ma>+8$⁠; [48]). The placement of these mafic units into the Kudz Ze Kayah formation requires that only those volcanic and volcaniclastic rocks that overlie the occurrences of mantle ultramafic rocks or are demonstrably part of the Fire Lake thrust sheet are included in the Fire Lake formation (e.g., boninite and IAT affinity; Figure 2).

Rocks in the Lower Mississippian Wolverine Lake group have previously not been differentiated into formations but described as basal, footwall, and hanging wall assemblages with respect to VMS mineralization [16, 30, 63, 64, 70]. We herein propose definitions for four informal “formations,” named after lakes and creeks present within the Wolverine Lake group based on lithological, geochemical, and geochronological characteristics presented in this paper and in previous work (e.g., [16, 17, 45]): (1) Little Jimmy formation, (2) Little Wolverine formation, (3) Go Creek formation, and (4) Jasper Creek formation (Figure 4(b)).

The Little Jimmy formation is the basal unit of the Wolverine Lake group and consists of mixed quartzofeldspathic-pebble conglomerate, grit, sandstone, and carbonaceous phyllite (Figure 4(b); [16]); previously, the rocks of this unit have been mapped as unit “U-5l” [62]. The Little Jimmy formation (ca. 357.6 to 356.4 Ma) is named after the Little Jimmy Lake, approximately 3.5 km directly south of the Wolverine mine site, and the type locality is immediately east of the regional unconformity between the Grass Lakes and Wolverine Lake groups (NAD83 9U 429446 mE, 6812106 mN; Figure 4(b)).

The Little Wolverine formation (ca. 356.4 to 355.0 Ma) stratigraphically overlies (and is locally intercalated with) the Little Jimmy formation and consists of felsic volcanic rocks, intercalated argillite, and high-level porphyritic intrusive rocks that host the ~6.2 Mt Wolverine Zn-Pb-Cu-Ag-Au deposit and other regional VMS prospects [16, 17, 71]. The Little Wolverine formation is named after Little Wolverine Lake that is immediately southeast of Wolverine Lake (Figure 4(b)). The rocks in this unit comprise what has previously been termed “footwall” or “U-5f/qfp,” “U-5cp,” and “U-6FW” [45, 62] and “Unit 1” within the Wolverine/Lynx zones [71].

The Go Creek formation stratigraphically overlies the Little Wolverine formation and is composed of rocks in the immediate hanging wall of the VMS deposits, including exhalites, iron formations, resedimented volcanic rocks, and carbonaceous argillite (Figure 4(b)). This unit is named after Go Creek, a watercourse that flows downhill through the Go Creek formation from the topographic highs in the Campbell Range basalts, joining with the Money Creek to the southeast (Figure 4(b)). Rocks of this unit have previously been classified as the “hanging wall” or “U-6HW” [45, 62] and include both “Unit 2” and “Unit 3” within the Wolverine/Lynx zones [71]. The formation includes a diverse group of rocks that are distinct in lithology and geochemistry from the underlying Little Wolverine formation (e.g., [45]).

Lastly, the Jasper Creek formation lies stratigraphically above the Go Creek formation and caps the Wolverine Lake group (Figure 4; [16]). The Jasper Creek formation contains massive to pillowed basalts and minor greywacke and carbonaceous argillite. This formation is named after Jasper Creek, which originates at high levels in topography north-northeast of the southern end of Wolverine Lake (Figure 4(b)). Previous work has placed this unit into “U-6HW” [62] or “Unit 4” [71] and correlated it with the Lower Permian Campbell Range formation of Slide Mountain terrane. Although the age of the Jasper Creek formation has not been directly determined isotopically or paleontologically, recent mapping leads us to interpret that the Jasper Creek formation is in depositional continuity with the underlying Go Creek formation and does not correlate with the Campbell Range Formation north of the Jules Creek fault [16]; therefore, unconformities no longer exist between lithologic contacts (Figures 3, 4(b), and 5).

Several bodies of mafic-ultramafic rocks have been mapped near rocks in the upper part of the Money Creek thrust sheet. These occurrences were originally interpreted as intrusions related to the Lower Permian Campbell Range Formation (i.e., Slide Mountain terrane) that subsequently pinned the timing of major thrust faults in the Finlayson Lake district [16]. Devine et al. [66], however, reinterpreted several of these mafic-ultramafic bodies as being klippen of the Cleaver Lake thrust sheet. We have adopted this interpretation and have reclassified mafic rocks originally mapped as intrusions of Early Permian Slide Mountain terrane to now being related to Late Devonian to Mississippian Yukon-Tanana terrane stratigraphy (see above regarding Jasper Creek formation; Figure 2). These discrete mafic-ultramafic bodies are interpreted to represent klippen of Devonian to Mississippian rocks that occur along an east-southeasterly trend between the Money Klippe and Klatsa metamorphic complex. As these rocks are no longer interpreted as intrusions, they provide no age constraints for the thrust faults that bound the Big Campbell, Money Creek, and Cleaver Lake thrust sheets; thrusting must have occurred post-deposition of the Money Creek formation but prior to intrusion of Early Cretaceous granitoids. Therefore, we propose that all three major thrust faults in the Finlayson Lake district can be reasonably interpreted as synthetic faults to the Jura-Cretaceous Inconnu thrust fault system that obducted Yukon-Tanana and Slide Mountain terranes onto the Laurentian continent (Figure 3).

Analytical procedures are outlined below and presented in detail in Supplementary Materials along with standard data and results from reproducibility monitoring (Tables S1, S4, and S5 and Figure S7). Sample preparation, mineral separation and extraction, annealing, and imaging for U-Pb zircon geochronology samples were carried out at Memorial University of Newfoundland (MUN; Figures S3–S5). Laser ablation- (LA-) ICP-MS and chemical abrasion- (CA-) ID-TIMS analyses were performed at the Isotope Geology Laboratory at Boise State University, Idaho (BSU). Laser ablation was performed on each grain using a New Wave Research UP-213 Nd:YAG UV laser (213 nm), where ablated material was analyzed on a ThermoElectron X-Series II quadrupole ICP-MS for U-Th-Pb isotopic ratios and trace element concentrations. Following LA-ICP-MS analysis, individual zircon grains were plucked from the epoxy mount and chemically abraded following the same procedure as in Manor et al. [30]. Isotopic determinations for CA-ID-TIMS analyses follow the methodology of Davydov et al. [72] and Schmitz and Davydov [73]. U-Pb ratios were measured using an IsotopX PhoeniX-62 thermal ionization mass spectrometer at BSU.

Sample preparation and measurement of major and trace element lithogeochemical data were performed at ALS Laboratories in North Vancouver, British Columbia, and Sudbury, Ontario; Geoscience Laboratories (Geo Labs) of the Ontario Geological Survey in Sudbury, Ontario, for transition metals, base metals, and semimetals; and Pacific Centre for Isotopic and Geochemical Research (PCIGR) at the University of British Columbia for trace elements on a subset of samples. At ALS, samples were fused with a lithium metaborate flux, digested in a HNO₃-HCl mixture, and analyzed by ICP-AES for major elements and ICP-MS for trace elements. At Geo Labs, samples were digested on hot plates using a four-acid mixture and analyzed on a Perkin-Elmer Elan 9000 ICP-MS following the methodology of Burnham [74]. At PCIGR, samples prefaced by 98DM-, P98-, P99-, and P00- [6, 17, 45, 61] were processed using high-pressure dissolution techniques [75] and analyzed on a Thermo Finnigan Element2 HR-ICP-MS.

Whole-rock Hf and Nd isotope ratios were measured at PCIGR using separate aliquots of sample powders used for trace element analyses. The Nd and Hf isotope ratios were measured by MC-ICP-MS following Weis et al. [76, 77] and normalized to the JNDi and JMC 475 standards for Nd and Hf isotopes, respectively. Analyses were normalized to JNDi with a $143Nd/144Nd=0.512116$ [78] and JMC 475 with a $176Hf/177Hf=0.282160$ [79]. All reference data presented as comparison fields have been renormalized to the accepted values of JNDi and JMC 475 as above. The present-day CHUR values used include $143Nd/144Nd=0.512638$ [80] and $176Hf/177Hf=0.282785$ [81], and parent-daughter ratios are $147Sm/144Nd=0.1967$ [82] and $176Lu/176Hf=0.0336$ [81]. Depleted mantle model ages (⁠$TDM$⁠) for Hf were calculated on using $176Hf/177Hf=0.283238$ and $176Lu/177Hf=0.03976$ [83]; for Nd, the $TDM$ ages were calculated using $143Nd/144Nd=0.513113$ and $147Sm/144Nd=0.2114$ [84].

Mineral liberation analysis (MLA) of polished thin sections of U-Pb TIMS samples was performed using a FEI Quanta 650 scanning electron microscope (SEM; CREAIT facility, MUN) equipped with MLA software. Each thin section was mapped to establish the modal abundances and varieties of minerals present in each rock. Quantitative modal abundances and false color thin section maps can be found in Table S6 and Figure S2, respectively.

The CA-ID-TIMS U-Th-Pb data for zircon in the Finlayson Lake district provide significantly greater resolution than U-Pb dates obtained by LA-ICP-MS methods (Tables 1 and 2). Results are presented as ²⁰⁶Pb/²³⁸U isotopic ratios for all the CA-ID-TIMS and LA-ICP-MS $samples<1$ Ga; those grains with $dates>1$ Ga are reported as ²⁰⁷Pb/²⁰⁶Pb dates. Presentation of the U-Pb dates follow the nomenclature of Schoene et al. [85] whereby the weighted mean U-Pb dates include uncertainties as $±X/Y/Z±2σ$⁠, where $X$ is the internal error without all systematic errors, $Y$ includes tracer calibration error, and $Z$ includes uncertainty related to both tracer calibration and decay constant error. Unless otherwise noted, uncertainties below are reported as $±X$ as the dates originate from the same laboratory. High-precision CA-ID-TIMS dates are used in the text below to aid interpretations related to the primary magmatic crystallization history and tectonic evolution of the district; the LA-ICP-MS results for CA-ID-TIMS samples are provided in Supplementary Materials. Detailed U-Th-Pb CA-ID-TIMS and LA-ICP-MS geochronology results are reported in Table 1 and Table S2, respectively, and a summary table compares the results of each method in Table 2. The CA-ID-TIMS and LA-ICP-MS U-Pb results are displayed as Wetherill concordia diagrams in Figures 8 and 9 and Figures S7 and S8. The CA-ID-TIMS data for the Kudz Ze Kayah formation and Wolverine Lake group (except sample 17MM-004) were previously reported in Manor et al. [30].

Zircon was separated from a felsic intrusion (17MM-047) that cuts mafic tuffs and lapilli tuffs in the hanging wall of the massive sulfide mineralization in the Kona VMS deposit (Figure S1; FL97-109; 204 m depth). Ten concordant zircon grains gave a LA-ICP-MS weighted mean ²⁰⁶Pb/²³⁸U date of $263.9±3.8$ Ma (2σ; $MSWD=3.1$⁠; Figure S7; Table S2), which is interpreted to be the crystallization age of this rock.

A basal grit (17MM-004) from the Little Jimmy formation was sampled from the same sample locality as the detrital sample of Murphy et al. [16], immediately northeast of the angular unconformity that separates the Grass Lakes and Wolverine Lake groups (Figures 4 and 6(a)). The CA-ID-TIMS results gave six concordant zircon fractions with a consistently low Th/U (0.2–0.4) and a weighted mean ²⁰⁶Pb/²³⁸U date of $357.658±0.096$ Ma (⁠$2σ$⁠; $MSWD=0.53$⁠; Figure 8(a); Table 1), which is inferred to represent the maximum depositional age of this rock.

(1) Grass Lakes Plutonic Suite. Rocks dated by U-Pb geochronology from the Grass Lakes plutonic suite include two granitoids (17MM-005 and 18MM-105) and two feldspar porphyritic rocks (17MM-061 and 17MM-062; Figures 7(a)–7(c) and Figure S1). A moderately deformed, medium-grained granodiorite (17MM-005) was collected immediately southwest of the angular unconformity that juxtaposes the Grass Lakes and Wolverine Lake groups (Figure 4). The CA-ID-TIMS data reveal six concordant fractions with $Th/U=0.3–0.5$ and a weighted mean ²⁰⁶Pb/²³⁸U date of $360.94±0.10$ Ma (⁠$2σ$⁠; $MSWD=1.3$⁠; Figure 8(b); Table 1), interpreted as the crystallization age of the rock. Sample 18MM-105 is a medium to coarse-grained granodiorite located south of the Kudz Ze Kayah formation in the North Lakes area and represents one from the most voluminous part of the Grass Lakes pluton (Figure 4). For CA-ID-TIMS, seven concordant zircon fractions give highly variable Th/U (0.3–2.3), similar to the North Lakes diorite. A weighted mean ²⁰⁶Pb/²³⁸U date of $361.801±0.099$ Ma (⁠$2σ$⁠; $MSWD=0.80$⁠; Figure 8(c); Table 1) is inferred to be the crystallization age of this granitoid.

A quartz-feldspar porphyritic intrusive rock (17MM-061) was sampled from a ~50 m thick intrusion in the upper 130 m of drill hole K16-372, which is interpreted to be near to the same stratigraphic level as the Kudz Ze Kayah VMS deposit. This sample contains medium-grained alkali feldspar phenocrysts and lesser quartz grains in a grey, glassy matrix (Figure 7(c)). Five of seven grains yielded concordant U-Th-Pb results, with Th/U between 0.5 and 1.3 and a weighted mean ²⁰⁶Pb/²³⁸U date of $361.71±0.13$ Ma (⁠$2σ$⁠; $MSWD=0.92$⁠; Figure 8(d); Table 1), interpreted as the crystallization age of this intrusive rock. Of the other two grains, one older grain (z3) gave an older ²⁰⁶Pb/²³⁸U date of $363.82±0.34$ Ma and is interpreted as a xenocryst; the other (z2) is significantly younger with a ²⁰⁶Pb/²³⁸U date of $349.03±0.32$ Ma but is much more discordant (~3%) relative to the other grains (<0.5%). This younger grain is interpreted to have not seen complete mitigation of Pb-loss through the chemical abrasion process. An additional feldspar-rich porphyritic intrusive rock (17MM-062) was sampled at the bottom of hole K16-372 (585 m depth). In contrast to 17MM-061, this rock is dominantly composed of plagioclase and quartz phenocrysts. Eight concordant zircon fractions yield moderate Th/U (0.6–0.8) and a weighted mean ²⁰⁶Pb/²³⁸U date of $361.901±0.086$ Ma (⁠$2σ$⁠; $MSWD=0.92$⁠; Figure 8(e); Table 1), which is interpreted to be the crystallization age of the sample.

(2) Wind Lake Formation. A felsic crystal tuff (17MM-001) was collected at an outcrop locality that displays interbedded felsic and mafic volcaniclastic rocks, ~500 m stratigraphically above the Kudz Ze Kayah–Wind Lake formation contact (Figures 4 and 6(e)). For CA-ID-TIMS, six concordant zircon fractions gave a tight range of Th/U (0.4–0.6) and a weighted mean ²⁰⁶Pb/²³⁸U date of $360.89±0.12$ Ma (⁠$2σ$⁠; $MSWD=0.47$⁠; Figure 8(f); Table 1), inferred as the eruption age for this rock. One zircon fraction yielded a ²⁰⁶Pb/²³⁸U date of $361.77±0.26$ Ma, which is beyond the upper limit of $2σ$ error of the weighted mean date and inferred to represent an antecryst or xenocryst.

(3) North Lakes Intrusion. Zircon from a diorite in the North Lakes intrusion (96DM-065; [16]) was reanalyzed with modern CA-ID-TIMS techniques for consistency in this study. The CA-ID-TIMS results gave six concordant zircon fractions with high Th/U (0.4–2.3) and a weighted mean ²⁰⁶Pb/²³⁸U date of $362.63±0.05$ Ma (⁠$2σ$⁠; $MSWD=1.6$⁠; Figure 8(g); Table 1). The new results are interpreted to be the crystallization age for this diorite and are outside of uncertainty from previously published results (weighted mean ²⁰⁷Pb/²⁰⁶Pb $date=365.0±1.2$ Ma; [16]).

Three granitoids from the Simpson Range plutonic suite were dated with LA-ICP-MS (P99-82, P99-24, and 18MM-107) and the latter two samples were dated by CA-ID-TIMS (Table 2, Figures 7(g) and S1). Sample P99-82 is a hornblende quartz diorite from the southeastern margin of the pluton (Figure 4). The LA-ICP-MS U-Th-Pb data reveal 35 concordant zircon fractions with a weighted mean ²⁰⁶Pb/²³⁸U date of $354.1±2.2$ Ma (⁠$2σ$⁠; $MSWD=1.3$⁠; Table S2), which is interpreted as the crystallization age of this granitoid. The hornblende granodiorite (P99-24) is located in the northeastern corner of the pluton (Figure 2). The CA-ID-TIMS results for this sample yielded five concordant fractions with Th/U tightly constrained to 0.5–0.6 and correspond to a weighted mean ²⁰⁶Pb/²³⁸U date of $352.12±0.10$ Ma (⁠$2σ$⁠; $MSWD=0.29$⁠; Figure 8(h); Table 1); this date is interpreted as the crystallization age of this rock. A monzogranite (18MM-107) was sampled near the core of the Simpson Range pluton, about 100 m south of the locality for 18MM-108 (Figure 4). At this location, lobate magma mingling textures are observed between monzogranite and gabbroic rocks. Seven concordant CA-ID-TIMS zircon fractions gave a limited range of Th/U (0.4–0.5) and a weighted mean ²⁰⁶Pb/²³⁸U date of $349.207±0.049$ Ma (⁠$2σ$⁠; $MSWD=0.77$⁠; Figure 8(i); Table 1), interpreted to be the age of crystallization for both this monzogranite and the coeval gabbroic rocks.

A quartz-porphyritic intrusive rock (18MM-108) was sampled from outcrop in the Cleaver Lake formation in the Cleaver Lake thrust sheet (Figure 7(h)), where a sharp contact with the Simpson Range granitoids is observed. The U-Th-Pb results gave eight concordant zircon fractions for CA-ID-TIMS, with moderate Th/U (0.4–0.6) and a weighted mean ²⁰⁶Pb/²³⁸U date of $363.185±0.058$ Ma (⁠$2σ$⁠; $MSWD=0.71$⁠; Figure 8(j); Table 1) that is interpreted to represent the crystallization age of this intrusive rock.

The lithogeochemical results from this study show significant element mobility due to hydrothermal alteration, notably in alkali (e.g., Na, K, and Ca) and large ion lithophile elements (LILE: Cs, Ba, Rb, K, Sr, and U; e.g., Figure 10(a)). However, several samples have $Al2O3/Na2O<10$ and Na₂O contents between 2 and 5 wt.%, which are indicative of relatively unaltered rocks (i.e., “least-altered” rocks) as observed in previous studies in the Finlayson Lake district (e.g., [45]). The alteration of least-altered felsic rocks, particularly the extrusive volcanic or volcaniclastic rocks of the Kudz Ze Kayah formation and Wolverine Lake group, resulted in the dominant sericite and alkali feldspar alteration assemblages with minor chlorite-pyrite overprint; the mafic rocks typically exhibit more abundant carbonate, epidote, and chlorite-pyrite alteration (Figure 11(a); [43, 86, 87]). The methodology outlined by Piercey et al. [45] is utilized here, where HFSE (Zr, Hf, Nb, Ta, and Y), Th, and REE (La to Lu) are assumed to be immobile, and their measured concentrations are indicative of primary petrologic processes rather than secondary alteration (e.g., [88]). Moreover, immobile elements and ratios that contain Al₂O₃, TiO₂, HFSE, and REE are used to assess the primary geochemical characteristics of the rocks (Figures 10–14). For additional comparison, felsic rocks have been normalized to upper continental crust (UCC; [89]), mafic rocks to primitive mantle (PM; [90]), and both felsic and mafic rocks to the C1 chondrite for REE (C; [91]). Complete lithogeochemical results are presented in Table S3.

Rocks were sampled from the drill core in the Kona VMS deposit including a single mafic tuff, intermediate-felsic clastic sedimentary rocks, and the Permian felsic intrusive rock dated in this study (Figures 2, 3, and S1). The mafic tuff has Nb/Y, Ti/V (<10), and Zr/Y ratios that are indicative of an island-arc tholeiitic to boninitic affinity (Figures 10(b), 10(d), and 10(f); e.g., [47, 48]), with a relatively flat primitive-mantle normalized pattern (⁠$La/YbPM=0.4,La/SmPM=1.2,andGd/LuPM=0.4$⁠) that overlaps with the boninite field (Figures 10(d) and 14(c)). Intermediate to felsic rocks consist of clastic sedimentary rocks (siltstones; $SiO2=62$–65 wt.%) and the Permian Si-rich felsic dike (⁠$SiO2=84wt.%$⁠) that occur proximal to massive sulfide mineralization in the Kona VMS deposit. Trace element ratios (Nb/Y and Zr/Y) indicate that these samples have calc-alkalic arc affinities (Figure 10(e)) and have similar REE abundances (⁠$La/YbUCC=0.7$–0.9) and neutral to positive Eu anomalies (Eu/Eu$∗$_UCC = 1.0–1.3; Figure 13(d)).

Felsic volcaniclastic $n=28$ and coherent volcanic and intrusive rocks $n=8$ and mafic intrusive rocks $n=5$ of the Kudz Ze Kayah formation were sampled (Figures 3–7; Table S3). The rocks show no geochemical preference to lithological variations in the felsic rocks (e.g., volcaniclastic vs. coherent; Figure 12(a)). The felsic rocks display a range of SiO₂ contents (59–78 wt.%) and have variable alkali concentrations due to hydrothermal alteration (Figures 10(a) and 11(a)). These rocks are subalkaline (⁠$Nb/Y=0.3$–0.7) and calc-alkalic (⁠$Zr/Y>4.2$⁠; [92]) to alkaline-peralkaline (⁠$Nb/Y=0.7$–1.1; $Zr>500$ ppm), both with intraplate A-type affinities (Figure 10). Two distinct groups are defined based on Zr/Al₂O₃, Zr/Ti, and Zr/Nb ratios: (1) group FN1, which includes alkaline and peralkaline dacite, trachyte/trachydacite, and minor rhyolite (e.g., TAS; Figure 10(a)), has within-plate Nb/Y ratios (Figure 10(c)), plots above the FI-FII rhyolite field (i.e., high La/Yb_C vs. Yb_C; Figure 12(b)), and has the highest Nb/Ta values (12–17, $mean=15$⁠; Figure 18), and (2) group FN2, which has predominantly subalkaline (to minor alkaline) affinities, falls within the andesite/basalt and trachyandesite fields (Nb/Y plot; Figure 10(b)), straddles the within-plate (A-type)–I-type boundary (Figure 10(c)), dominantly plots in the FII (±FI and FIIIa) rhyolite fields (e.g., [34]), and has lower Nb/Ta than group FN1 (10–14, $mean=12$⁠; Figure 18). Upper continental crust-normalized immobile elements yield relatively similar patterns between rocks of groups FN1 and FN2, except for higher absolute trace element abundances in group FN1. The patterns are flat with no significant LREE-enrichment (La/Sm_UCC: $FN1=0.8$–1.1; $FN2=0.8$–1.1), relatively higher HREE (Gd/Lu_UCC: $FN1=0.6$–1.8, $mean=1.4$⁠; $FN2=0.8$–1.9, $mean=1.1$⁠), and variably negative Eu (Eu/Eu$∗$_UCC: $FN1=0.4$–0.9, $mean=0.7$⁠; $FN2=0.2$–1.3, $mean=0.5$⁠) and strong negative Ti (Ti/Ti$∗$_UCC: $FN1=0.05–0.09;FN2=0.04–0.18$⁠) anomalies (Figure 13(c)). These results are consistent with A-type volcanic signatures [45], except for samples 18MM-142 (tuff with minor argillite) and 18MM-143 (high-level felsic intrusive rock), which show elevated metal concentrations (Ni, Cu, Zn, Cr, and V; Figure 13(c); Table S3), and 18MM-106 (biotite-rich felsic tuff); all three of these samples have lower La/Yb_UCC than typical Kudz Ze Kayah formation rocks (Figure 13(c)).

Mafic rocks sampled from the Kudz Ze Kayah formation comprise (1) fine-grained, subvolcanic intrusive rocks within the area of the Kudz Ze Kayah deposit (⁠$n=3$⁠) and (2) medium to coarse-grained intrusive rocks (metagabbro) in outcrop stratigraphically below the Kudz Ze Kayah formation (⁠$n=2$⁠; Figure 4(a)). The subvolcanic intrusive rocks plot within the subalkaline basalt field (⁠$Nb/Y=0.4$–0.5) and have MORB/back-arc basin (BABB) affinities (Ti/$V=30$–36; Figure 10(d); [93]). The Zr/Y (~4.5) also indicate a transitional to calc-alkalic, E-MORB-like (Th/Nb ~0.1–0.2; Figure 11(d)) affinity. The gabbroic rocks, however, show relatively higher Nb/Y (1.3–1.6) and Ti/V (43–52) ratios indicative of an alkalic affinity (Figures 10(b) and 10(d)); results for these samples plot near the OIB field (Figures 11(d) and 11(e)). The primitive mantle-normalized diagrams for the subvolcanic intrusive rocks show relatively shallow patterns with slightly higher LREE over HREE (⁠$La/YbPM=3$–4), near-neutral Eu and slight negative Nb abundances (⁠$Eu/Eu∗PM=0.88–0.96;Nb/Nb∗PM=0.19–0.24$⁠). Patterns for gabbroic rocks are steeper (⁠$La/YbPM=5–34$⁠) and have higher Nb anomalies (⁠$Nb/Nb∗PM=0.5–1.9$⁠; Figure 14(b)). Both lithologies show geochemical similarities to other Nb-enriched basalts to E-MORB-type mafic rocks in the district (Figure 14(b); [48, 53, 61]).

Mafic rocks from the Wind Lake formation have basalt and alkaline basalt to basaltic trachyandesite compositions (Figures 10(a) and 10(b)) and MORB/BABB to alkalic geochemical affinities (⁠$Ti/V=30$–160; Figure 10(d)) and straddle the subalkaline to alkaline boundary with $Nb/Y=0.42$–2.3 (Figure 10(b)). A pillow basalt sample (17MM-054) gave higher Nb/Y, which falls within the alkaline field, and has elevated Nb/Yb and Th/Yb and plots over the non-arc mantle array; the rest of the mafic rocks are subalkaline and plot above the mantle array (Figure 11(d)). Primitive mantle-normalized diagrams show steep patterns for the mafic intrusions and tuffs with high La/Yb_PM (4.5–10) and negative Eu and Nb troughs $Eu/Eu∗PM=0.6–1.0;Nb/Nb∗PM=0.06–0.17$⁠; the pillow basalt (17MM-054) has a steeper pattern with a slight positive Eu and more elevated Nb abundance (⁠$La/YbPM=21;Eu/Eu∗PM=1.1;Nb/Nb∗PM=0.3$⁠; Figure 14(a)). These are similar to non-arc basalts that have experienced variable crustal contamination (Figure 14(a); [61]). The felsic tuff sample (17MM-001) has higher SiO₂ (74 wt.%) and geochemical characteristics of group FN1 rocks with a trachytic composition (Figure 10(b)) and has an alkalic and peralkaline, within-plate geochemical affinity (⁠$Nb/Y=0.76;Zr/Y=14.2;Zr/Nb=125$⁠; Figures 10 and 11). The upper continental crust-normalized pattern is relatively flat (⁠$La/YbUCC=0.86$⁠) and overlaps with other within-plate, A-type felsic volcanic rocks in the district (Figure 13(b)).

One gabbro (18MC-015) from the North Lakes intrusion has an alkaline basalt composition (⁠$Nb/Y=0.9$⁠) and plots on the boundary between MORB/BABB and alkalic fields (⁠$Ti/V=50$⁠; Figure 10(d)). The sample yields Th-Nb compositions that fall on the boundary of the non-arc mantle array (Figure 11(d)), between other non-arc mafic rocks in the district. The primitive mantle-normalized signature has a shallow slope with LREE-enrichment relative to HREE $La/YbPM=7.4$⁠, slightly positive Eu $Eu/Eu∗PM=1.1$⁠, and negative Nb $Nb/Nb∗PM=0.2$ anomalies, indicating these rocks have Nb-enriched basalt affinities (Figure 14(b)).

Samples of the Grass Lakes plutonic suite come from a batholith-sized intrusion and high-level felsic porphyritic intrusive rocks (Figures 2 and 4). Samples from the granitic pluton have high SiO₂ (68–79 wt.%) and HFSE concentrations with within-plate signatures that are transitional to calc-alkaline (⁠$Nb/Y=0.52–0.70;Zr/Y=2.9–14$⁠; Figures 10(b), 10(c), and 10(e)) to slightly alkalic (⁠$Nb/Y=0.70–0.80$⁠). The suite contains rocks with both FN1 and FN2 group characteristics similar to the Kudz Ze Kayah formation, where FN1 rocks have lower SiO₂ (<73 wt.%) and higher Zr (>390 ppm), Zr/Al₂O₃, Zr/Ti, and Zr/Nb than group FN2 rocks. Upper continental crust-normalized diagrams show distinct trends for the pluton (⁠$La/YbUCC=0.64–0.92$⁠) and feldspar porphyritic dikes (⁠$La/YbUCC=1.1–1.8$⁠); the dikes have FN1 affinities and contain slightly greater abundances of LREE than group FN2 granitoids. The HFSE and REE abundances for both groups overlap with rocks of the Kudz Ze Kayah formation, Grass Lakes plutonic suite, and Wolverine Lake group footwall (e.g., $Eu/Eu∗UCC=0.4–0.9;Ti/Ti∗UCC0.03–0.1$⁠; Figures 13(b) and 13(c); [6, 45]).

The Wolverine Lake group samples (⁠$n=12$⁠) consist of a lower quartzofeldspathic grit (17MM-004) in the Little Jimmy formation and volcaniclastic rocks (⁠$n=7$⁠), a quartz-feldspar porphyritic intrusive rock (QFP), and feldspar porphyritic intrusive rocks (FP; $n=3$⁠) from the Little Wolverine formation. Two distinct groupings comprise the geochemical signatures of these rocks, similar to the Kudz Ze Kayah formation: (1) group FN1, which contains only FP intrusive rocks in the upper footwall nearest to VMS mineralization, and (2) group FN2, which contains the basal grit, volcaniclastic, and intrusive rocks throughout the Little Jimmy and Little Wolverine formations. Group FN1 rocks have moderate to high SiO₂ (58–77 wt.%), peralkaline to alkaline trachyte and rhyolite with high Zr/Y (>10), within-plate (⁠$Nb/Y=0.9$–1.1), and mostly FI rhyolite affinities (Figure 12(b)) and have high Nb/Ta (15–17.5). Group FN2 rocks have high SiO₂ (75–80 wt.%), fall in the andesite/basalt, trachyandesite, and trachyte fields and contain both within-plate A-type and I-type signatures (⁠$Nb/Y=0.5$–1.1; Figure 10(b)), and have tholeiitic to calc-alkaline Zr/Y (2.5–11) and FI, FII, and FIIIa affinities and lower Nb/Ta than rocks in group FN1 (10–13). The upper continental crust-normalized patterns for group FN1 show more abundant REE relative to group FN2 (La/Yb_UCC: $FN1=1.8–2.5,mean=2.1;FN2=0.5–1.5,mean=1.0$⁠), with relatively similar, flat LREE (⁠$La/SmUCC=1.0–1.2$ vs. 0.7–1.2, respectively) and steeper HREE slopes (⁠$Gd/LuUCC=1.3–2.3,mean=1.8$⁠) than in group FN2 (⁠$Gd/LuUCC=0.7–1.1,mean=1.1$⁠; Figure 13(a)). Negative Ti and Eu anomalies are similar in group FN1 and FN2 rocks, where $Ti/Ti∗UCC=0.05$–0.06 and 0.02–0.07, and Eu/Eu$∗$_UCC = 0.5–0.6 and 0.2–1.0, respectively (Figure 13(a)).

Four samples from the Go Creek formation have distinct geochemical signatures (FN3; e.g., [46] Figures 10–12). The rocks are typically strongly quartz-altered (>83 wt.% SiO₂) that are classified as andesite to basalt (Figure 10(b)), with transitional to calc-alkalic (⁠$Zr/Y=4$–12) and volcanic arc, I-type to M-type affinities (⁠$Nb/Y<0.6$⁠; Figure 10(b)). They have low Zr/Ti and Zr/Nb (i.e., $Zr<110$ ppm) and chondrite-normalized REE (e.g., La/Yb_C vs. Yb_C) signatures that are different than the Little Jimmy and Little Wolverine formations and have arc-like signatures similar to volcanic rocks from the Cleaver Lake thrust sheet (see text below; Figure 13(a)). Upper continental crust-normalized immobile element signatures show flat patterns with slightly positive slopes, but with more depleted abundances compared to the Little Wolverine formation $La/YbUCC=0.4–1.3$⁠. The patterns show neutral to positive Eu anomalies $Eu/Eu∗UCC=0.9–1.3,mean=1.1$ and less negative Ti (⁠$Ti/Ti∗UCC=0.08$–0.18, $mean=0.13$⁠) that contrast with those lower values from the footwall (Figure 13(a)).

Felsic flows and porphyritic intrusive rocks $n=3$ from the Cleaver Lake formation are andesite to rhyolite; two samples (18MM-108 and P99-45) plot as calc-alkalic rhyolites and straddle the subalkaline-alkaline boundary as rhyolite/dacite to trachyandesite (⁠$Nb/Y=0.7–0.9$⁠), and the third (P99-39) plots as a transitional andesite with a much lower Nb/Y (0.2; Figure 10(b)). The calc-alkalic rhyolites have volcanic arc-like Nb-Y signatures with I-type affinities, whereas the andesite plots in the ocean-ridge (OR-type) field (Figure 10(c)). The upper continental crust-normalized patterns show distinct trends for the andesite, which shows a positive slope with greater HREE relative to LREE $La/YbUCC=0.3$⁠, compared to the calc-alkalic rhyolites $La/YbUCC=1.0–1.3$⁠, yet HREE show similar, flat patterns $Gd/LuUCC=~0.9$ where the transitional rocks have more abundant HREE (Figure 13(e)). All three of these rocks have a prominent negative Ti trough $Ti/Ti∗UCC=0.04–0.09$⁠, but the calc-alkalic rocks have negative Eu anomalies compared to a slight positive signature in the transitional rock (⁠$Eu/Eu∗UCC=0.8$–1.0 vs. 1.1, respectively). These signatures overlap with those of calc-alkalic to tholeiitic arc rocks from both the Cleaver Lake and Fire Lake formations [45].

One mafic augite porphyritic rock (18MM-109) has a basaltic composition that plots as a transitional island arc tholeiite (IAT; $Nb/Y=0.3;Zr/Y=3.7;Ti/V=18$⁠; Figure 10). The rock sample has a Th-Nb composition that plots above the non-arc mantle array, interpreted to indicate contributions from a subducting slab in an arc environment (Figure 11(d)). The primitive mantle-normalized diagrams show elevated LREE over HREE (⁠$La/YbPM=4.0$⁠) and distinct negative Nb and Ti (⁠$Nb/Nb∗PM=0.1;Ti/Ti∗PM=0.1$⁠) and slightly negative Eu anomalies (⁠$Eu/Eu∗PM=0.86$⁠; Figure 14(d)), characteristic of IAT affinity rocks (e.g., [48]).

A hornblende granodiorite (P99-24) and monzogranite (18MM-107) from the Simpson Range plutonic suite have moderate SiO₂ contents (65–69 wt.%) with subalkaline affinities (⁠$Nb/Y=~0.7$⁠; Figure 10(b)). The rocks are calc-alkalic (⁠$Zr/Y=8$–11), fall within the volcanic arc (I-type) field, and typically have Zr $contents<180$ ppm and $Nb<11$ ppm that correspond to low Zr/Ti (0.03–0.04) and high Zr/Nb (12–16). The upper continental crust-normalized diagrams have relatively flat patterns (⁠$La/YbUCC=1.1–1.5$⁠) with slight negative Nb and Ti (⁠$Nb/Nb∗UCC=~0.2$⁠; $Ti/Ti∗UCC=~0.2$⁠) and positive Eu anomalies (⁠$Eu/Eu∗UCC=1.0–1.3$⁠; Figure 13(f)).

Whole-rock Hf and Nd isotopic compositions (⁠$n=28$⁠) were measured for rocks in the Fire Lake, Big Campbell, and Cleaver Lake thrust sheets from the Finlayson Lake district (Tables 3 and 4; Figure 15). The results below are presented by thrust sheet, then in order of stratigraphically lowest to highest. In the Fire Lake thrust sheet, a Permian felsic dike that crosscuts mafic volcaniclastic rocks in the Kona VMS deposit hanging wall gave $εHfiandεNdi$ of –4.7 and –7.3, respectively. In the Grass Lakes group, a mafic intrusive rock in the lowest stratigraphic levels of the Kudz Ze Kayah formation (17MM-056; previously mapped as Fire Lake formation) has a juvenile isotopic composition (⁠$εHfi=+1.5;εNdi=+2.4$⁠). Felsic rocks from the Kudz Ze Kayah formation (⁠$n=8$⁠) gave a range of evolved isotopic compositions of $εHfi=–13.7to–2.8andεNdi=–11.2$ to –4.5. Rocks in the immediate hanging wall and footwall of VMS deposits (e.g., Kudz Ze Kayah and GP4F) have more chondritic isotopic compositions (⁠$εHfi=–7.0to–2.9;εNdi=–7.3$ to –4.9) relative to those rocks highest in the hanging wall (17MM-002), where $εHfiandεNdi$ are –13.7 and –11.2, respectively. A mafic subvolcanic intrusive rock in the footwall of the Kudz Ze Kayah deposit (17MM-060) gave near-chondritic $εHfi=+2.7andεNdi=–0.5$⁠. Rocks of the Wind Lake formation were sampled either as felsic (17MM-001) or mafic (17MM-003) volcaniclastic rocks. The range of isotope values constrained to near-chondritic relative to the Kudz Ze Kayah formation, yielding $εHfi$ from –1.6 to +2.7 and $εNdi$ from –3.7 to –0.5, where the felsic tuff (17MM-001) has the most negative value. Granitoids from the Grass Lakes plutonic suite (⁠$n=5$⁠) give the largest range of evolved isotopic compositions across the district (⁠$εHfi=–15.7to–3.0;εNdi=–11.4$ to –3.9).

In the Wolverine Lake group, the basal quartzofeldspathic grit (17MM-004; Little Jimmy formation) yielded the most evolved isotopic compositions of the unit (⁠$εHfi=–15.8;εNdi=–11.3$⁠), which broadly overlap with rocks in the Kudz Ze Kayah formation and Grass Lakes plutonic suite. Four samples from the Little Wolverine formation give restricted, evolved isotope compositions (⁠$εHfi=–8.1to–6.0;εNdi=–8.5$ to –7.0) that overlap with feldspar porphyritic rhyolites (⁠$n=2$⁠) from the Wolverine/Lynx zone (⁠$εHfi=–7.3;εNdi=–7.9$⁠) and the Fisher zone (⁠$εHfi=–7.9;εNdi=–7.8$⁠).

Three samples from the Cleaver Lake thrust sheet include felsic volcanic rocks of the Cleaver Lake formation $n=2$ and a hornblende granodiorite from the Simpson Range plutonic suite $n=1$⁠. The Hf and Nd isotopic compositions for the Cleaver Lake formation are generally chondritic and correspond to the geochemical affinity of the rocks. Sample P99-39 is of tholeiitic affinity and yields juvenile isotope compositions $εHfi=+4.8;εNdi=+0.25$ relative to P99-45, which is of calc-alkaline affinity and gives evolved values $εHfi=–4.0;εNdi=–4.8$⁠. The Simpson Range plutonic suite granodiorite yields a much more evolved isotopic signature where $εHfiandεNdi$ are –17.1 and –13.1, respectively.

Felsic and mafic rocks in the Finlayson Lake district have been subdivided into distinct groups based on their immobile major and trace element concentrations and isotopic compositions (Figure 12). The characteristics of each group are defined by least-altered, precursor rocks with distinct trends in immobile ratio-ratio diagrams (e.g., Figure 12). These groupings are also discussed in terms of stratigraphic packages in the Finlayson Lake district. There are multiple precursor rocks within each rock package, but for the scope of this paper, the details of the causes of these multiple populations are not fully evaluated.

In the Big Campbell thrust sheet, rocks of non-arc affinity (i.e., back-arc) are divided into felsic groups FN1, FN2, and FN3 and mafic group MN1 (Figure 12). Groups FN1 and FN2 comprise rocks of non-arc affinity (i.e., back-arc) and contain most volcanic and plutonic rocks in the Grass Lakes and Wolverine Lake groups (Figure 12). Group FN1 felsic rocks have high HFSE (e.g., $Zr>360$ ppm) and REE, low Al₂O₃, similar TiO₂, and less evolved εHf and εNd (>–8) compared to the FN2 group (εHf and $εNd<–8$⁠). The FN3 group, however, contains Go Creek formation rocks that have arc-related geochemical affinities, despite being located immediately above FN1 and FN2 rocks in non-arc stratigraphy (Figures 4 and 5); these signatures have been attributed to either derivation from extrabasinal arc sources or significant crustal contamination of non-arc magmas [7, 45]. The group FN3 rocks have the lowest Zr (<150 ppm), HFSE, and REE contents of all felsic rocks in the district, Zr/Al₂O₃ like group FN2 rocks, distinct Th/Nb, but low Zr/Ti ratios most similar to arc-related felsic rocks in the Cleaver Lake thrust sheet. Group MN1 contains all mafic rocks in the non-arc environment that are characterized by high Ti/V (>20) and Zr/Y (>4; e.g., N-MORB, E-MORB, and OIB; Figures 10(d) and 10(f)).

In the Cleaver Lake thrust sheet, felsic rocks of arc affinity are split into groups FA1 and FA2, which are typically designated as the Cleaver Lake formation (ca. 363 Ma) and Simpson Range plutonic suite (ca. 354–349 Ma), respectively. Group FA1 felsic rocks have higher Zr, REE, Zr/Al₂O₃, and Zr/Ti, lower TiO₂, and less evolved εHf and $εNd>–5$ than group FA2. Arc-related mafic rocks in the Cleaver Lake formation comprise group MA1, which includes MORB to IAT-type mafic rocks that have similar Ti/V and Zr/Y to the non-arc mafic rocks.

In the Fire Lake thrust sheet, mafic rocks of arc affinity are separated into group MA2 (Figures 10(d) and 10(f)). This group contains mafic rocks of the Fire Lake formation that have boninite and IAT geochemical affinities that are characterized by low Ti/V (<20) and Zr/Y (<3; Figures 10(d) and 10(f)).

MLA-SEM imaging was completed on thin sections of back-arc and arc rocks in the Finlayson Lake district to test mineral abundances relative to geochemical results. All samples show varying modal abundances of HFSE-REE-bearing minerals that account for <0.7 vol.% of the total rock composition (Table 5; Table S6). Rocks in all groups contain between 0.13 and 0.31% apatite, except for the lowest abundances in 17MM-001 (0.05%), 18MM-102 (0%), and 18MM-108 (<0.01%). Group FN1 rocks contain greater abundances of rutile (0.07-0.29%), zircon (0.04-0.08%), and monazite (0.02-0.05%) relative to the FN2 group. Rocks in group FN2 contain between 0.03 and 0.06% titanite and allanite, whereas these minerals are <0.01% in group FN1 rocks. Xenotime, uraninite, and thorite are typically present as trace constituents below 0.01%. Two arc samples (group FA2) from the Cleaver Lake thrust sheet were analyzed, one monzogranite (18MM-107) and one quartz-porphyritic rhyolite (18MM-108). Sample 18MM-107 has high titanite (0.38%) and allanite (0.06%), with comparable zircon (0.03%) and apatite (0.19%) abundances to group FN2 rocks; all other minerals are traced below 0.01%. Sample 18MM-108, however, contains 0.02% rutile but trace abundances of all other accessory minerals (Table 5).

The geochemical and isotopic signatures of mafic rocks in the Yukon-Tanana terrane provide critical insight into the tectonomagmatic environment of formation, and relative roles that enriched lithospheric mantle, depleted asthenospheric mantle, slab metasomatism, and crustal contamination played in their genesis (Figure 11; [7, 61]). These various components will be evaluated below.

Deciphering the mantle component (i.e., enriched or depleted) in ancient mafic rocks requires utilization of elements (and isotopes) that are immobile, insensitive to fractional crystallization, and reflective of the mantle source regions and crustal contamination and/or slab metasomatism (e.g., [94, 95]). The immobile element ratios of Th/Yb, Nb/Yb, and TiO₂/Yb are useful for delineating mantle sources as these ratios are generally reflective of incompatible enrichment of a mantle source (i.e., MORB-OIB array in Figures 11(d) and 11(e); [94]); however, the Th/Yb ratio is also sensitive to crustal contamination and/or slab fluid components as they typically have higher Th/Yb ratios at a given Nb/Yb ratio and distinctive trajectories in Th/Yb-Nb/Yb space (Figure 11(d); [61, 94–97]). The mafic rocks found in the back-arc region of the Finlayson Lake district that plot predominantly within the MORB-OIB array in Th/Yb–Nb/Yb space (Figure 11(d)) are interpreted to be derived from variably enriched mantle sources. For example, rocks from the lower Kudz Ze Kayah formation have relatively depleted BABB (~N-MORB) signatures (e.g., [48]), the upper Kudz Ze Kayah formation have weakly enriched (~E-MORB) signatures, and the Wind Lake formation has highly enriched (~OIB-like) signatures; metagabbros of the North Lakes intrusion and the lower Kudz Ze Kayah formation (samples 17MM-056 and 17MM-057) plot between E-MORB and OIB (Figure 11(d)). Mafic rocks from the arc regions, however, include IAT affinity rocks of the Cleaver Lake formation and boninite to IAT signature rocks of the Fire Lake formation that plot above the MORB-OIB array proximal to E-MORB and N-MORB, respectively (Figure 11(d)). The variation in incompatible element enrichment between these various rocks potentially reflects mixing and varying contributions from depleted, N-MORB-like asthenospheric melts and enriched lithospheric mantle-derived melts (e.g., [7]). The Hf-Nd isotope results for these rocks support this hypothesis. In particular, the back-arc mafic rocks typically have $εHf350Ma=+0.8$ to +2.6 and $εNd350Ma=–0.8$ to +2.2 (Figures 12 and 14), which, in the case of Nd data, overlaps with the limited dataset of Piercey et al. [61] (Figure 15); however, several E-MORB basalts in the lowest stratigraphic levels of the Kudz Ze Kayah formation overlap with juvenile $εNd=+8.1$ [48]. Many of the samples with both near-chondritic and juvenile εHf and εNd have flat to positive Nb anomalies and the greatest incompatible element enrichment, indicative of rocks that have not seen slab metasomatism nor crustal contamination, suggesting that these isotopic values are a function of the incompatible element-enriched mantle source they were derived from. It is therefore suggested that (1) the near-chondrite Hf-Nd isotope signatures were generated from a mixture of depleted asthenospheric mantle and a component of preexisting and enriched lithospheric mantle [7, 48, 53] and (2) the juvenile Nd isotope signatures, particularly in the E-MORB samples (c.f., [48]), were generated from a nearly uncontaminated asthenospheric mantle source and erupted following little to no residence time in the lithospheric mantle. These geochemical and isotopic signatures are therefore distinct from other mafic rocks in the Finlayson Lake district that experienced crustal contamination, as evidenced by high Th/Yb and more evolved isotopic signatures (Figures 11(d)–11(f), and 15).

A subset of mafic rocks in the belt displays distinct geochemical ratios with $Nb/LaPM<1$⁠, $Nb/ThPM<1$⁠, and elevated Th/Yb (Figures 12 and 14) indicative of either a subducting slab input or crustal contamination (Figure 11; [7]). Two boninite and IAT basalt samples from the Fire Lake and Cleaver Lake formations, respectively, have negative Nb anomalies and high Th/Nb, which are interpreted to be generated from the retention of Nb and enrichment of Th due to fluid mobilization from the subducting slab (e.g., Figures 11(d), 11(f), 14(c), and 14(d); [45, 48, 62, 95, 96]). These samples also overlap with rocks that have high εNd (+3 to +7; [48]) and are interpreted to have formed from primitive arc magmas that have not interacted with evolved continental crust. Further, the rocks with BABB signatures in the Kudz Ze Kayah formation also show negative Nb anomalies, slightly elevated Th/Yb, and high $εNd=+8.5$ (Figures 11(d) and 11(f); [48]), which is interpreted to indicate that back-arc affinity magmas were overprinted by a weak subduction signature. In contrast, other mafic rocks with back-arc affinities have high Th/Nb and Th/Yb, lower εHf and εNd (–1 to +3), and 1.0-1.4 Ga model ages (Tables 3 and 4), which suggested that the mafic melts interacted with continental crust (e.g., [61]).

These data illustrate that the mafic rocks in the Finlayson Lake district reflect varying contributions of asthenospheric mantle and continental lithospheric mantle that were overprinted by subducting slab or evolved continental crust components. These results also suggest that different crustal architectures in both the arc and back-arc environment generated the variable isotopic and trace element compositions of these mafic rocks. Back-arc mafic rocks, the exception of the lowest Kudz Ze Kayah formation (i.e., BABB and E-MORB), interacted with old subcontinental lithospheric mantle and/or continental crust during ascent, whereas arc rocks did not. Therefore, some components of evolved lithosphere should be expected during the formation of felsic rocks in the back-arc, which is explored below.

Felsic rocks throughout the Finlayson Lake district have distinct geochemical and isotopic signatures that define them as arc or back-arc rocks (e.g., Nb/Y, Zr/Y, Zr/Ti, and SiO₂) and reflect varying contributions from evolved crust and juvenile mantle-derived components (e.g., Nb/Ta ~9-18; $εHf350Ma=–17to+4;εNd350Ma=–13$ to +0.2; Figures 10, 15, and 16; [6, 7, 17, 45, 48, 61, 70]). Arc rocks from the Cleaver Lake thrust sheet have highly variable isotopic values (⁠$εHf350Ma=–17$ to +4; $εNd350Ma=–13$ to +0.2) and show distinct crystallization ages, including (1) calc-alkalic to tholeiitic volcanic rocks of the Late Devonian Cleaver Lake formation (⁠$εHf350Ma=–4$ to +5; $εNd350Ma=–5$ to +0.2) and (2) calc-alkalic granitoids of the Early Mississippian Simpson Range plutonic suite (⁠$εHf350Ma=–17;εNd350Ma=–13$⁠; Figure 15). These suites of arc-related felsic rocks have a range in HFSE and REE contents from very depleted (tholeiitic rocks) to less depleted but significantly lower concentrations than the back-arc felsic rocks (e.g., Zr and $REEsum<200$ ppm; Figures 11 and 16), while having isotopic signatures that reflect influence from the variable crustal basement during Late Devonian and Early Mississippian magmatism. Back-arc rocks, however, illustrate a significant decoupling of Hf-Nd isotopes and HFSE-REE contents, where these signatures define two geochemical groups that underwent differing crustal melting processes (Figures 15–17). There is a positive relationship between εHf and εNd and HFSE-REE contents with the VMS-proximal FN1 suite rocks having the higher values when compared to the VMS-distal FN2 rocks. These features are attributed to the variable contributions of mantle and crustal materials, degrees of crustal assimilation, and temperatures of melting and emplacement and are explored further below.

The back-arc geochemical groups, FN1 and FN2, have isotopic and geochemical features that show evidence for contributions from both mantle and crustal sources; however, the degree of crustal melting and contributions from underplated basaltic magmas is variable (Figure 16). The variations in the basaltic/juvenile component are best recorded by (1) distinct Nb/Ta ratios (Figure 16), (2) bulk rock Hf-Nd isotopic compositions (Figure 15), and (3) Th/U in zircon (Figure 9(b)). The Nb/Ta values for group FN1 are ~17 (i.e., mantle-like), whereas FN2 values are ~12 (i.e., crustal-like; [17, 89, 98, 99]), and these also correspond with εHf and εNd values that are most juvenile (i.e., closest to chondrite—group FN1) and most evolved (group FN2; Figure 15), respectively. In zircon crystals, the highest Th/U (>1) are associated with samples that have εHf-εNd greater than ~–8 and Nb/Ta ~17 (e.g., sample 17MM-031, group FN1). However, some samples have very wide ranges in zircon Th/U in a single sample, such as in plutonic samples 96DM-065 and 18MM-105 that gave large ranges of Th/U between 0.4-2.3 and 0.3-2.2, respectively (Figure 9(b)). These variations could reflect mixing between juvenile and evolved magmas in the chamber; however, equally viable is that it reflects potentially varying degrees of disequilibrium crystallization where zircon crystallization is not controlled by specific partition coefficients in the melt (e.g., [100]). Even with the variability in Th/U in some samples, there is a general trend at both the whole rock and mineral scale of more juvenile and higher temperature magmatic rocks proximal to VMS mineralization (i.e., the FN1 population).

Isotopic mixing models indicate that rocks from the Finlayson Lake district formed from variable mixtures of depleted N-MORB mantle melts and crustal melts from the assimilation of preexisting Snowcap assemblage rocks (Figures 15, 17, and 18). In the Cleaver Lake thrust sheet, arc rocks in the Cleaver Lake formation reveal between 9 and 21% Nd and Hf crustal components compared to 48-57% for the Simpson Range plutonic suite. In the Big Campbell thrust sheet, back-arc mafic rocks have isotopic compositions that indicate ~8-13% Nd and Hf contributions from the Snowcap endmember, whereas the felsic rocks vary from 17 to 45% Hf and 19 to 47% Nd from this endmember (Figure 15; Tables 3 and 4). The group FN1 rocks have the lowest contributions from the crustal endmember (17-28% Hf and 19-33% Nd), which also corresponds to the most elevated εHf-εNd, whereas group FN2 rocks have greater crustal contributions (24-45% Hf and 28-47% Nd) that correlate with the most evolved isotopic values (Figure 15). Therefore, it is evident that crust-mantle mixing was a dominant process in the formation of felsic rocks in the Finlayson Lake district and was particularly important in the differences in petrogenetic histories of the group FN1 and FN2 rocks.

The whole-rock lithogeochemistry, petrogenesis, and tectonic setting of VMS-bearing felsic rocks have been extensively studied [34, 36, 43, 116, 117]. Similarly, U-Pb geochronology has provided constraints on the timing of magmatism and mineralization in some VMS districts globally [28, 92, 118–121]. Moreover, there have been studies of shallow subvolcanic intrusions with close spatial and/or temporal ties to VMS-hosting stratigraphy, and these have also been implicated as the primary source of heat for the formation of VMS deposits [31–34, 122]. However, several authors have challenged this hypothesis and argue that these intrusions may not have been the direct heat source for mineralization but rather passive magmatic products in a regional, thermally anomalous geodynamic environment [17, 35, 36].

In the Finlayson Lake district, the Grass Lakes plutonic suite was previously suggested to be the heat driver for mineralization in the Kudz Ze Kayah formation [6]. The new U-Pb dates for the Grass Lakes plutonic suite (ca. 361.8 to 360.9 Ma; Figures 8 and 9) indicate that the intrusive rocks postdate the Kudz Ze Kayah VMS deposit and volcanism and therefore could not be the heat source for VMS mineralization. Additionally, while there is not a large plutonic suite associated with mineralization in the Wolverine Lake group, there are high-level syn-VMS porphyritic intrusions proximal to the Wolverine deposit that have been linked to regional extensional-related magmatism [17]. Despite their variations in timing, these rocks have geochemical and isotopic signatures that suggest potential fertility for VMS mineralization. For example, the FN1 rocks within <100-200 m of the ca. 362.8 Ma Kudz Ze Kayah deposit, the ca. 355.2 Ma Wolverine deposit, and the ca. 354.9 Ma Fisher zone have elevated HFSE-REE concentrations (e.g., $Nb/Ta>~15$⁠), zircon saturation temperatures (>890°C), and εHf-εNd isotope signatures that are consistent with formation at high-temperatures related to basaltic underplating (Figures 15, 16, and 19). In contrast, VMS-barren FN2 group rocks have distinct U-Pb dates outside of 2s uncertainty from the FN1 rocks, are HFSE-REE poor, and have more evolved εHf-εNd isotope signatures. The spatial association of VMS with magmas indicative of rifting and with evidence of high temperatures of emplacement and more juvenile signatures suggests that basaltic underplating association with back-arc rifting controlled the localization of VMS mineralization in the Finlayson Lake district. Rifting provided the ground conditions (i.e., synvolcanic faults) needed to focus fluid flow (e.g., [123]), whereas the basaltic underplate provided the essential heat needed to drive hydrothermal fluid circulation that formed both large and high-grade VMS deposits in the district (e.g., [17]). The underplating of rifts by basaltic magmas and associated crustal melting are processes likely recorded in many VMS districts globally (e.g., [36]), and further study of both bulk rock and mineral scale chemistry and isotopes is required to test these ideas and their validity.

Most magmatic rocks that comprise the Finlayson Lake district formed between ca. 363.5 and 354.8 Ma (Figure 9; [30]) and were part of an east-dipping Japan-Japan Sea style arc and back-arc environment along the western Laurentian continental margin (e.g., [2, 6, 7, 15, 124]). This model has been scrutinized by van Staal et al. [50] who argued that the tectonic configuration of the Yukon-Tanana terrane was more like the New Britain-Solomon-New Hebrides arc system at the Australian-Pacific plate boundary (e.g., [125]). Our results allow testing of these models and will utilize the (1) stratigraphic evaluation of mafic rocks in the Fire Lake and Kudz Ze Kayah formations and (2) geological characteristics of the ca. 360-358 Ma deformational event recorded in the Finlayson Lake district. An outcome of our work will be a series of testable models that will help guide future studies reconstructing the potential Late Devonian tectonic configuration of the Yukon-Tanana terrane and related rocks along the western Laurentian margin.

Mafic magmatism in the Fire Lake formation was previously interpreted to represent arc–back-arc rift initiation on the western Laurentian continental margin that subsequently evolved into a regime dominated by crustal melting and within-plate magmatism in the Kudz Ze Kayah formation [16, 45, 48]. This interpretation relied on three pieces of evidence: (1) ca. 365 Ma dates for mafic intrusions interpreted to be comagmatic with the Fire Lake formation (i.e., North Lakes intrusion; [4, 16]), (2) geochemistry of mafic rocks previously interpreted to represent the initiation and evolution of rifted continental arc (e.g., arc boninite and IAT to non-arc MORB to OIB; [47, 48]), and (3) apparent stratigraphic continuity of mafic rocks of the Fire Lake formation and felsic rocks of the Kudz Ze Kayah formation [16]. We will reevaluate the above in light of recent ideas (i.e., [50]) and new results and geological relationships.

Most mafic rocks in the Kudz Ze Kayah formation have chondritic isotopic compositions (εNd and εHf ~0) that have been interpreted as being mixtures of asthenospheric and enriched lithospheric mantle sources (±crustal/slab influence; see discussion above). The products of this magmatism were predominantly rocks with the most enriched E-MORB to OIB affinities; however, several mafic rocks stratigraphically overlie the North River formation that has primitive BABB to E-MORB affinities with juvenile $εNdi$ values (+8.1 and +8.5; [48]). These isotopic compositions are significantly more juvenile than the results from the rest of the redefined Kudz Ze Kayah formation and, given their stratigraphic position, are interpreted to be one of the first products related to rift-related magmatism in the nascent continental back-arc basin. At this stage, early rift-related magmas were likely generated by high-temperature mantle melting where the depleted asthenospheric melts (e.g., BABB) interacted with metasomatized mantle from the mantle wedge as the back-arc opened (e.g., [95, 126]). These primitive melts also contained an enriched E-MORB component that ascended through extensional structures in the continental crust with insufficient residence time in enriched lithospheric mantle and continental crust to influence the isotopic compositions. As rifting continued, relatively deeper, incipient mantle melts would have had increased opportunity to interact with enriched lithospheric mantle and generated weakly alkalic rocks stratigraphically higher in the Kudz Ze Kayah formation (e.g., [94]), which then progressed stratigraphically upwards into a bimodal but felsic-dominated, within-plate volcanic sequence that was subsequently intruded by the North Lakes intrusion (ca. 362.6 to 361.8 Ma) and the Grass Lakes plutonic suite (ca. 361.9 to 360.9 Ma). These observations are interpreted to suggest that the Kudz Ze Kayah formation, as opposed to the Fire Lake formation (c.f., [48]), represents the pre-363.3 Ma marker of bimodal magmatism and back-arc rift initiation in the Yukon-Tanana terrane.

In the redefined Fire Lake formation, the allochthonous ultramafic rocks are overlain by extrusive metavolcanic rocks with affinities common to primitive arcs (e.g., boninite and IAT; [48, 68]). Such associations have been found in modern arcs and in many ophiolites and interpreted to have formed due to rapid rifting and exhumation of ultramafic rocks, as proposed for suprasubduction zone forearc [127–129] and back-arc environments [130–132]. Using examples from modern arcs, we herein propose a revised petrogenetic and geodynamic setting for the redefined Fire Lake formation and associated ultramafic slab. Ultramafic-mafic rocks and high-Ca boninites were generated in the forearc of the nascent arc system following subduction initiation, where rapid extension eventually led to a transfer of extensional stresses from the forearc to the back-arc region (e.g., [128]). Forearc extension and/or incipient back-arc formation thus facilitated deep exhumation of ultramafic mantle rocks in this model. Depleted boninites erupted above the ultramafic-mafic rocks, formed via decompression melting, due to rising asthenospheric melts that passed through a depleted, refractory mantle wedge that had previously extracted melt in the arc proper (e.g., IAT; [48, 133]). These magmas would have depleted the mantle wedge prior to boninite formation and have left an ultradepleted residue that the boninitic rocks could have been derived from (e.g., [134]). The proposed forearc setting provides the most ideal scenario to explain the preservation of oceanic rocks in the Fire Lake formation and the juxtaposition of these rocks near continental affinity rocks, which requires that the Fire Lake formation and associated ultramafic rocks (i.e., the arc initiation suite) are thrust onto the inboard North River formation.

The proposed model for the Fire Lake formation implies that continental crust was absent in the formation of most rocks, in contrast to the suggestions of Piercey et al. [47] who suggested the boninites formed in an environment with stratigraphic continuity with continental crust (e.g., a Havre Trough-type setting in the Kermadec arc; [135, 136]). The new observations and data presented here are not internally consistent with this model, and we suggest the Fire Lake formation formed in a separate oceanic domain proximal but outboard of the continental crust-floored Kudz Ze Kayah formation back-arc and were juxtaposed by thrusting. These observations require stratigraphic redefinition of both geological units and imply that the Kudz Ze Kayah formation, not the Fire Lake formation, records the onset of continental arc–back-arc rifting on the western Laurentian margin in the Late Devonian (pre-363.3 Ma).

The ca. 360-358 Ma deformation, uplift, and erosion are recorded by the angular unconformity between the Grass Lakes and Wolverine Lake groups [16, 62]; the cause of the deformation and unconformity is poorly understood but broadly correlates with timing of the Late Devonian-Early Mississippian Antler Orogeny in the western United States (e.g., Figure 20; [137–140]). Evidence to support the nature and age of this unconformity includes the following: (1) quartzofeldspathic grits in the Little Jimmy formation have immature textures with angular feldspar laths, interpreted to be representative of deposition that was coincident with or shortly after seafloor volcanism; (2) zircon in this sample that yielded a maximum depositional age of ca. 357.6 Ma indicates deposition coeval with magmatism in the Wolverine Lake group; (3) the unit has geochemical signatures identical to the underlying Kudz Ze Kayah and overlying Little Wolverine formations but has a very evolved isotopic signature in group FN2 (see above); (4) rocks below the unconformity are more highly deformed than those above, having well-developed cleavages, ductile fabrics, and locally exhibit isoclinal folding; (5) relatively undeformed granitoids cut folded granitic dikes, and both were subsequently uplifted and eroded; and (6) the unconformity cuts kilometers of stratigraphy observed as a beveled and tilted erosional feature from the Wind Lake formation down to the North River formation (Figure 2).

New constraints on volcanism in the Wind Lake formation and the new CA-ID-TIMS date of ca. 357.6 Ma for the grit sample (Figures 8(a), 8(f), and 9(a)) indicate that the deformation period occurred between ca. 360.9 Ma and ca. 357.6 Ma. A similar style of unconformity has been recognized in the Money Creek thrust sheet between the Waters Creek and Tuchitua formations, but regional correlations beyond the Finlayson Lake district to other parts of the Yukon-Tanana terrane or North American continental margin are lacking. We interpret this to indicate that deformation at ca. 358 Ma was not more widely recognized in the northern Cordillera due to poor preservation of the back-arc environments in the rock record. Further, this deformation is present during a period of sinistral transpression on the western Laurentian margin [57, 137, 141–146], suggesting that it represents rarely preserved evidence that could be related to the Antler Orogeny in the northern North American Cordillera (Figure 20). If this is the case, the Antler Orogeny would have impacted the entire western North American continental margin during the Late Devonian to Early Mississippian (e.g., [140]), but much of the evidence in the northern Cordillera was then variably erased during younger deformation and metamorphic events.

Despite there being an unconformity between the Grass Lakes and Wolverine Lake groups, its timing suggests it may be unrelated to deformation in the region but due to the evolution of the subsequent rift-drift stage (e.g., [147]). Two types of Grass Lakes granitoids are present immediately beneath the unconformity: (1) strongly deformed granite that preserves isoclinal folds interpreted to correlate with a transpressional deformation event (i.e., Antler Orogeny) and (2) a later, weakly foliated granite that crosscuts the strongly deformed granites above. The latter granitoid intrusion was dated at $357.3±2.8$ Ma [16] and, despite the large uncertainty, overlaps with the maximum depositional age and new CA-ID-TIMS date in the grits above the unconformity (Figure 8(a)). The younger cross-cutting granite is located below the unconformity and constrains the upper age limit of deformation, but it also establishes a distinct, post-transpressional magmatic event that preceded uplift and erosion. The ca. 357 Ma granite could also provide a local source for the Little Jimmy formation grits. Therefore, it appears that the erosional unconformity may be slightly younger and thus unrelated to the observed ca. 360-358 Ma transpressional deformation event.

As a result of the “decoupled” nature of deformation and erosion, we propose that the basal Wolverine Lake group unconformity represents a rift-onset unconformity that quickly merged into a breakup unconformity, marking the outboard flank of the Slide Mountain ocean. Early rifting in the Grass Lakes group transitioned upwards into alkalic magmatism but was then aborted [7], then was succeeded by compressional deformation. Post-deformation extension quickly facilitated an abrupt, pre-rift uplift period on the rift flank of the new Wolverine Lake group basin as a regional isostatic response to upwelling mantle [147–149]. The rapid change in topography on the continental back-arc resulted in an erosional surface (i.e., the basal Wolverine Lake group unconformity) that cut kilometers of stratigraphy and deposited sediment into the new low-lying basin between ca. 357.6 and 356.4 Ma (i.e., Little Jimmy formation; Figures 4(b) and 9; e.g., [30]). Continental margin extension in the Wolverine Lake group back-arc then resumed for at least ~1.5 Myr before MORB-type basalts of the Jasper Creek formation were erupted. The Jasper Creek formation includes both E-MORB and N-MORB basalts with $εNdi$ between –4 and +7 [57], which has been interpreted to suggest that Jasper Creek formation basalts have both lithospheric and asthenospheric mantle components [57]. The presence of lithospheric components suggests a linkage to the continental lithosphere, and the asthenospheric component implies incipient back-arc spreading. Together with overlying stratigraphic and structural relationships, it suggests that the entire Wolverine Lake group stratigraphic section represents the evolution of crustal breakup, rifting, and the rift-drift transition following lithospheric rupture leading to incipient spreading (Figures 4 and 5; [45, 53, 71]). Further, these relationships fit with regional models for the Yukon-Tanana and Slide Mountain terranes and the development of a back-arc to ocean spreading center on the western Laurentian continental margin. Our results illustrate, however, this Devonian-Mississippian evolution was much more complex than previously recognized (e.g., [7, 16, 48, 124]).

Yukon-Tanana terrane rocks in the Finlayson Lake district, Yukon, represent early manifestations of mid-Paleozoic arc–back-arc magmatism immediately adjacent to the Laurentian continental margin which led to a rift-drift transition and the onset of the opening of the Slide Mountain ocean. When interpreted in the context of new structural and stratigraphic interpretations, our integrated geochronological, geochemical, and isotopic results indicate that the back-arc rocks formed during back-arc rifting in the Late Devonian to Early Mississippian (ca. 363.4 to 355.0 Ma), punctuated by a complex period of compressional deformation, postcompressional magmatism, uplift, and erosion between ca. 360 and 358 Ma. The Cleaver Lake arc contains coeval Late Devonian volcanic rocks (ca. 363.2 Ma) and multiple pulses of granitic plutonism between 352 and 348 Ma. The rocks in both the Grass Lakes and Wolverine Lake groups were generated above the continental crust of Laurentian affinity, which we interpret as rocks similar to the Snowcap assemblage that makes up the basement to Yukon-Tanana terrane throughout Yukon, Alaska, and British Columbia. Geochemical and isotopic results from felsic rocks in both rock packages indicate that crustal melting of the Snowcap assemblage occurred at different temperatures and correlates with variable degrees of assimilation. High-temperature (890-980°C) crustal melting facilitated low degrees of assimilation for Nd and Hf (<~35%) that generated rocks with group FN1 geochemical signatures (e.g., εHf and $εNd>–8$⁠); this group contains rocks proximal to VMS mineralization. Low-temperature signatures (780-850°C), however, correspond to increased assimilation (>40%), evolved isotopic signatures (εHf and $εNd<–8$⁠), and group FN2 geochemistry. These geochemical distinctions suggest that rocks directly hosting mineralization in the Kudz Ze Kayah and Wolverine VMS deposits were formed from high-temperature crustal melting from an increasingly mantle-derived component, interpreted to result from extensive basaltic underplating during an active period of slab rollback and mantle upwelling.

The conditions favorable for VMS generation were achieved through a series of complex tectonomagmatic configurations beginning in the Late Devonian. We propose that rapid and complex subduction initiation of various oceanic and continental crust blocks, which originated from remnants of an attenuated Laurentian margin, facilitated rapid arc mobility and extension in the back-arc regions. Rapid extension and magmatism in the Kudz Ze Kayah and Wolverine back-arc basins were punctuated by a period of transpressional deformation between ca. 360 and 358 Ma, possibly related to a rarely preserved segment of the Antler Orogeny. We propose that the deformation was separate from later uplift and erosion that define the regional angular unconformity between the two units, an unconformity interpreted as a rift-onset and overlapping breakup unconformity that marks the outboard flank and beginning of the continental margin rifting that led to the rift-drift transition and formation of the Slide Mountain ocean. This study highlights the importance of new, high-precision CA-ID-TIMS U-Pb zircon geochronology and its utility in placing tectonomagmatic processes within tight temporal chronostratigraphic frameworks. Further work in the Finlayson Lake district has the potential to define the earliest Late Devonian history (i.e., pre-Kudz Ze Kayah back-arc) of the Yukon-Tanana terrane and its relationship with the Laurentian continental margin, which can then be translated across the entire North American Cordilleran margin.

All of the data presented in this paper is available from the tables, figures, and supplementary files.

The authors are not aware of any conflicts of interest related to this research.

We acknowledge BMC Minerals and the Yukon Geological Survey for their ongoing support of this project. The new property-scale mapping, stratigraphic relationships, and sampling would not have been possible without tremendous assistance during fieldwork from Robin Black, Neil Martin, and Robert Burke (BMC Minerals) and Dillon Hume, Darcy Baker, Mark Baknes, Roger Hulstein, Ty Maggie, Ron Voordouw, and Victoria Tweedie (Equity Exploration). Jim Crowley, Mark Schmitz, and Craig Hart are thanked for their assistance with U-Pb geochronology, sample collection, organization, and logistics. Wanda Aylward and John M. Hanchar are thanked for their assistance with SEM work and mineral separations and preparation. Maurice Colpron helped with logistics, sample collection, and insightful discussions. Nikola Denisová, Rosie Cobbett, Adam Wiest, and Carly Mueller are thanked for edits on early versions of this manuscript and discussions of Cordilleran tectonics, VMS deposits, and geochemistry. We also thank Equity Exploration, Trans North helicopters, and Tintina Air for their support and transport while conducting field research from 2017 to 2019. Funding for this research was provided by the Yukon Geological Survey, BMC Minerals, NSERC Discovery Grant, an NSERC Collaborative Research and Development Grant, and the Targeted Geoscience Initiative 5 (TGI-5) program of the Geological Survey of Canada to S.J. Piercey and a GSA Graduate Student Research Grant and SEG Canada Foundation Student Research Grant to M.J. Manor.