Shear zones are repositories of the knowledge of crustal evolution in accretionary orogens. The Yukon-Tanana terrane of the North American Cordillera is a composite peri-cratonic terrane comprising segments with distinct protoliths and/or tectono-metamorphic histories separated by shear zones. In northwestern British Columbia, the Wann River shear zone separates the Carboniferous and older Florence Range suite from the pre-Carboniferous Boundary Ranges suite. Based on sheared early Jurassic plutons and post-metamorphic cooling ages, the shear zone has been previously interpreted to have been active between 185 and 170 Ma and subsequently folded. We investigated the similar to 4 km thick Wann River shear zone in two locations: (1) a klippe on White Moose Mountain where it is sub-horizontal and (2) north of Llewellyn Glacier similar to 40 km to the SSE where it is sub-vertical. Although initial field mapping was consistent with previous interpretations of a top-to-the-SE/sinistral sense of shear, quartz crystallographic axis preferred orientations (CPO) analyses and SHRIMP U-Pb geochronology on variably deformed intrusive rocks reveal a significantly more complex deformation history. We documented three deformation events along the Wann River shear zone. D-1 is a high-temperature (similar to 590-710 degrees C) top-to-the-SE (sinistral after folding) event constrained to <= 191 +/- 2 Ma. It is associated with crustal thickening and burial of the metamorphic suites. D-2 is a moderate-temperature (similar to 430-540 degrees C, possibly underestimated) dextral event constrained between 184 +/- 2 and 181 +/- 2 Ma and likely occurred during exhumation. D-3 is a low-temperature (similar to 315 degrees C) sinistral event constrained to <= 181 +/- 2 Ma. Our study illustrates how detailed microstructural characterization of a shear zone combined with geochronology can reveal its polyphase evolution.

Sulphide minerals within xenoliths provide critical records for melting and metasomatism in the lithospheric mantle and crust. During the transport of these mantle and crustal xenoliths to the surface, high-temperature sulphide minerals cool and transform to a low-temperature assemblage that includes pyrrhotite, pentlandite, and chalcopyrite. However, the small size and complex intergrowths between these low-temperature minerals present analytical challenges for determining their trace element concentrations, and for estimating the bulk sulphide composition prior to cooling. Herein we combine geochemical mapping by laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) with machine learning for a suite of basalt-hosted peridotite, pyroxenite, gabbro, and anorthosite xenoliths from British Columbia, Canada, to reconstruct the bulk sulphide composition and constrain the petrology of the Cordilleran lithospheric mantle. Sulphide mineral compositions extracted from the segmented geochemical maps for the peridotite and pyroxenite xenoliths yield chondritic to sub-chondritic Os/Ir and Pd/Ir ratios, indicating that these samples represent mantle residues that underwent low to moderate degrees of partial melting. However, reconstructed bulk sulphide compositions also yield suprachondritic Re/Os ratios and high concentrations of relatively incompatible and highly chalcophile trace elements (e.g., Te, Ag, and Se). These enriched geochemical signatures indicate residual sulphide phases must have mixed with a metasomatic and compositionally fractionated sulphide melt. Small volumes of alkaline silicate melts partially preserved with sulphide, feldspar, and other metasomatic minerals in the same micro-textural settings further indicate that sulphide metasomatism occurred in the mantle and prior to volcanism. New two-pyroxene and olivine-spinel geothermometry results demonstrate that these sulphide-bearing lherzolite, websterite, and wehrlite xenoliths equilibrated at a range of temperatures (931-1129 degrees C). Sulphide minerals with similar compositions from peridotite xenolith localities across British Columbia of varying equilibration temperatures and volcanic ages (i.e., Oligocene to Pleistocene) suggests that sulphide metasomatism is a characteristic feature of the Cordilleran lithospheric mantle.

Abstract Dashwoods is a composite peri-Laurentian terrane in Newfoundland and forms the basement to the Early Ordovician to Silurian Notre Dame arc. The southern part of Dashwoods is characterized by paragneiss that is intruded by Early Ordovician to Late Silurian plutons and affected by polyphase Taconic to Salinic deformation and high-grade metamorphism. The crystalline basement of Dashwoods is not exposed and pre-Middle Ordovician paragneiss is investigated herein to constrain the provenance of Dashwoods. SHRIMP U–Pb zircon analysis of the paragneiss yielded metamorphic rims ranging from c. 500 to 395 Ma and abundant detrital grain cores ranging from c. 1853 to 546 Ma. The presence of abundant Tonian dates differentiates Dashwoods from the adjacent Humber Margin in Newfoundland, and Hebridean and Grampian terranes in the British Isles. The detrital provenance of Dashwoods is most similar to the Baie Verte Margin in Newfoundland, and Tyrone Complex and Dalradian Supergroup in Ireland. These data suggest that Dashwoods and the Baie Verte Margin originated near the Rockall promontory and were subsequently emplaced outboard of the Humber Margin by Ordovician to Carboniferous motion along the Baie Verte–Brompton Line.

Keppie et al. (2021) presented a Devonian to Carboniferous oroclinal bending hypothesis and dextral strike-slip translation of a single arc terrane to explain the distribution of Cambrian to Ordovician volcanic rocks in the northern Appalachians.

The Neoproterozoic to Cambrian rifting history of Laurentia resulted in hyperextension along large segments of its Paleozoic margins, which created a complex paleogeography that included isolated continental fragments and exhumed continental lithospheric mantle. This peri-Laurentian paleogeography had a profound effect on the duration and nature of the Paleozoic collisional history and associated magmatism of Laurentia. During the initial collisions, peri-Laurentia was situated in a lower-plate setting, and there was commonly a significant time lag between the entrance of the leading edge of peri-Laurentia crust in the trench and the arrival of the trailing, coherent Laurentian landmass. The final Cambrian assembly of Gondwana was followed by a global plate reorganization that resulted in Cambrian (515–505 Ma) subduction initiation outboard of Laurentia, West Gondwana, and Baltica. Accretion of infant and mature intraoceanic arc terranes along the Appalachian-Caledonian margin of the Iapetus Ocean started at the end of the Cambrian during the Taconic-Grampian orogenic cycle and continued until the ca. 430–426 Ma onset of the Scandian-Salinic collision between Laurentia and Baltica, Ganderia, and East Avalonia, which created the Laurussian continent and closed nearly all vestiges of the Iapetus Ocean. Closure of the Iapetus Ocean in the Appalachians was followed by the Devonian Acadian and Neoacadian orogenic cycles, which were due to dextral oblique accretion of West Avalonia, Meguma, and the Suwannee terranes following the Pridolian to Lochkovian closure of the Acadian seaway and subsequent outboard subduction of the Rheic Ocean beneath Laurentia. Continued underthrusting of Baltica and Avalonia beneath Laurentia during the Devonian indicates that convergence continued between Laurentia and Baltica and Avalonia, which, at least in part, may have been related to the motions of Laurentia relative to its converging elements. Cambrian to Ordovician subduction zones formed earlier in the oceanic realm between Laurentia and Baltica and started to enter the Arctic realm of Laurentia by the Late Ordovician, which resulted in sinistral oblique interaction of the Franklinian margin with encroaching terranes of peri-Laurentian, intra-oceanic, and Baltican provenance. Any intervening seaways were closed during the Middle to Late Devonian Ellesmerian orogeny. Exotic terranes such as Pearya and Arctic Alaska became stranded in the Arctic realm of Laurentia, while other terranes such as Alexander and Eastern Klamath were translated further into the Panthalassa Ocean. The Middle/ Late Devonian to Mississippian Antler orogeny along the Cordilleran margin of Laurentia records the first interaction with an outboard arc terrane built upon a composite block preserved in the Northern Sierra and Eastern Klamath terranes. The Carboniferous–Permian Alleghanian-Ouachita orogenic cycle was due to closure of the vestiges of the Rheic Ocean and assembly of Pangea. The narrow, continental transform margin of the Ouachita embayment of southern Laurentia had escaped accretion by outboard terranes until the Mississippian, when it collided with an outboard arc terrane.

The Laberge Group is an Early to Middle Jurassic sequence of mostly siliciclastic sedimentary rocks that were deposited in a marginal marine environment in the northern Canadian Cordillera. It forms a long narrow belt with a total thickness of 3–4 km extending for more than 600 km across southern Yukon and northwestern British Columbia. These sedimentary rocks overlap the Yukon-Tanana, Stikinia and Cache Creek terranes that form the main components of the Intermontane superterrane. The Laberge Group contains a record of the erosion of some of these terranes, and also offers some constraints on the timing of their amalgamation and accretion to the Laurentian margin. The Laberge Group was deposited with local unconformity on the Late Triassic Stuhini Group (in British Columbia) and correlative Lewes River Group (in Yukon), both of which are volcanic-rich, and assigned to the Stikinia terrane. The Laberge Group is in turn overlain by Middle Jurassic to Cretaceous clastic rocks, including the Bowser Lake Group in BC and the Tantalus Formation in Yukon. Clast compositions and detrital zircon populations within the Laberge Group and between it and these bounding units indicate major shifts in depositional environment, basin extent and detrital sources from Late Triassic to Late Jurassic. During the Early Jurassic clast compositions in the Laberge Group shifted from sediment- and volcanic-dominated to plutonic-dominated, and detrital zircon populations are dominated by grains that yield ages that approach or overlap their inferred depositional ages. This pattern is consistent with progressive dissection and unroofing of (an) active arc(s) to eventually expose Triassic to Jurassic plutonic suites. Detrital rutile and muscovite data from the Laberge Group indicate rapid cooling and then exhumation of adjoining metamorphic rocks during the Early Jurassic, allowing these to contribute detritus on a more local scale. The most likely source for such metamorphic detritus is within the Yukon-Tanana terrane, and its presence in the Laberge Group may constrain the timing of amalgamation and accretion of the Yukon-Tanana and Stikinia terranes. Thermochronological data also provide new insights into the evolution of the Laberge Group basin. Results from the U–Th/(He) method on detrital apatite suggest that most areas experienced post-depositional heating to 60°C or more, whereas U–Th/(He) results from detrital zircon show that heating to more than 200°C occurred on a more local scale. In detail, Laberge Group cooling and exhumation was at least in part structurally controlled, with more strongly heated areas situated in the footwall of an important regional fault system. The thermochronological data are preliminary, but they suggest potential to eventually constrain the kinematics and timing of inversion across the Laberge Group basin and may also have implications for its energy prospectivity. In summary, the Laberge Group is a complex package of sedimentary rocks developed in an active, evolving tectonic realm, and many questions remain about the details of its sources and evolution. Nevertheless, the available information demonstrates the potential of combined geochronological and thermochronological methods applied to detrital minerals to unravel links between regional tectonics, basin development and clastic sedimentation.

West and East Ganderia in the northern Appalachians and Caledonides, respectively, represent a Gondwanan superterrane situated along the Tornquist margin of Amazonia prior to Furongian drift into the Iapetus Ocean, which opened the Rheic Ocean from west to east. The ocean-facing Penobscot arcbackarc system was established by 515 Ma in West Ganderia. A correlative arc formed at ca. 480 Ma in East Ganderia. In West Ganderia, the Tremadocian Penobscottian orogeny involved closure of the Penobscot backarc basin. Tremadocian Monian tectonism in East Ganderia was mainly related to oblique accretion to East Avalonia and the Megumian Cymru terrane. Penobscottian and late Floian Monian orogenesis led to termination of Early Ordovician arc magmatism, probably due to shallow subduction of buoyant oceanic lithosphere. Early to Middle Ordovician arc-backarc systems were erected on Penobscottian-Monian modified West and East Ganderia. The active edge of West Ganderia accreted diachronously to peri-Laurentia between 475 Ma and 455 Ma, followed by Wenlock to Ludlow Salinic accretion of the inboard Gander margin through closure of the intervening backarc basin. In the Caledonides, East Ganderia and East Avalonia accreted to Laurentia during the correlative Wenlock Scandian orogeny. The Ordovician to Silurian tectonic evolution of Ganderia was markedly non-cylindrical with pronounced partitioning of SalinicScandian convergence. Pridoli to Lochkovian closure of the Acadian seaway in the northern Appalachians led to Acadian accretion of West Avalonia to composite Laurentia. Shallow Early Devonian underthrusting of West and East Avalonia beneath Laurentia produced widespread Acadian tectonism and voluminous Early Devonian Acadian magmatism. The Appalachian Meguma terrane formed part of Megumia, which probably formed originally adjacent to East Avalonia and West Africa. The Meguma terrane accreted dextrally to Laurentia during and after the late Emsian to Famennian Neoacadian orogeny, mainly driven by outboard subduction of the Rheic Ocean. No correlative terrane docking took place in the Caledonides. Crown Copyright (c) 2021 Published by Elsevier B.V. on behalf of International Association for Gondwana Research. All rights reserved.

Garnet crystallization has been simulated in the MnNCKFMASHT model system using a simple nucleation and growth scenario, calibrated with three‐dimensional garnet crystal size distribution data as well as garnet compositional data obtained by electron probe micro‐analysis and laser ablation inductively coupled plasma mass spectrometry. Results indicate wide‐spread Barrovian‐type metamorphism for garnet‐zone rocks from the Snowcap assemblage along a hairpin‐shaped pressure–temperature loop with garnet growth from ~515°C and 4 kbar to metamorphic peak conditions of ~600°C and 6 kbar. Lu–Hf garnet‐whole geochronology points to initial garnet growth at c. 192.2 ± 4.7 Ma. Sm–Nd garnet–whole‐rock geochronology applied to a sample with garnet rims enriched in Sm indicates that the metamorphic peak conditions have been attained at c. 172.9 ± 2.4 Ma. Older garnet growth at c. 245.3 ± 0.8 Ma during a low‐P–high‐T event has been preserved as garnet cores separated from the Jurassic garnet rims by a sharp microstructural and compositional discontinuity. These polyphase garnets are restricted to Mn‐rich metapelitic lithologies. Trace element zoning in the outermost ~50 μm thin segments of the Early Triassic garnet cores reflects a short garnet growth episode in the presence of melt at peak conditions of ~710°C and 2.5 kbar, supported by phase equilibrium and diffusion geospeedometry calculations. Diffusion simulations across the interface between the Early Triassic garnet core and the Jurassic garnet rim indicate that the Barrovian‐type metamorphism during the Jurassic lasted for 20–25 Myr, in line with the radiometric data.

Keynote perspectives Cordilleran Geoscience 2020: Past, present and (?) future Jim Monger, Geological Survey of Canada, emeritus Establishing the distribution of rocks in space and time in the Canadian Cordillera began in 1871 when British Columbia entered Canadian Confederation. By 1906, a generalized geological map of the entire continent shows blank spaces in western Canada only in northern Yukon and North West Territories. Mapping is on-going and aided by increasingly refined tools for locating, dating and analysing rocks. Mapping led to attempts to understand the evolution of the mountain belts. By ~1875, Appalachian studies led to the “Geosynclinal Cycle” concept, in which strata deposited in long-lived troughs called geosynclines were buried, deformed, metamorphosed, intruded and uplifted, a hypothesis that dominated North American tectonics for ~90 years. During the 1950s-60s, geophysical mapping and sea-floor drilling of the ~70% of Earth’s surface beneath the oceans led by 1963 to formulation of the New Global Tectonics, aka Plate Tectonics. A GSA Penrose Conference on “The meaning of the New Global Tectonics for Geologists” was held in Asilomar, California in November 1969. In January 1970 its content was conveyed to Vancouver’s geological community at the first meeting of GAC’s Cordilleran Section. Some 200 participants were expected; ~800 attended, mostly from the mineral industry. The plate tectonic paradigm had an immediate influence. Geosynclinal rocks in the western Cordillera comprising associated volcanic and sedimentary strata were reinterpreted as magmatic arc and ocean floor deposits, and sedimentary successions in the eastern Cordillera as former continental margin deposits. The greatest challenge, readily apparent from global maps of ocean floor ages, was the potential within a region such as western Canada where the record of nearly 400 million years of plate convergence creates enormous paleogeographic uncertainty. That this is so was supported in the early 1970s by recognition of anomalous (w.r.t. the North American craton) paleomagnetic results and fossils, and led to the concept of “terranes”, which are regions with internally consistent geology whose original relations to one another and the continental margin are paleographically “suspect”. Today, we recognize that most terranes are defined by their youngest assemblage; e.g. Quesnellia in southern British Columbia is defined by early Mesozoic arc-dominated rocks that overlie late Paleozoic arc, ocean floor and possibly older craton margin strata. Re-examination of linkages between terranes suggests that all major terranes probably were assembled along the former continental margin of North America by the Middle Jurassic (≥174 Ma). There are two current “end-member” models of the principal controls on Cordilleran mountain-building. In one, plates are moved by “classic” ridge-push, trench-pull plate forces, and orogenesis is caused by arc-continent collision resulting from subduction dipping away from the continent. An alternative model proposes that plates containing cratons, like the North American Plate, are driven mainly by coupling between mantle and deep lithospheric continental roots and driven by mantle flow which results in deformation of weak, warm lithosphere of the leading plate margin. Resolution of these models calls for careful re-evaluation of the evidence for arc-facing directions and paleogeography. Playing the Great Game of Cordilleran terrane tectonics: A personal account 1970-2020 JoAnne Nelson, British Columbia Geological Survey, emeritus A half a century ago, plate tectonics grabbed the microphone in Cordilleran tectonic discussions and has never dropped it since. A key marker of the oncoming revolution was Tanya Atwater’s 1970 paper, which invoked oceanic plate motion as the major driver of the on-land tectonic evolution of western North America, notably motion on the San Andreas fault, but by implication, everything. I am a child of that revolution. I inhaled the Atwater paper as a third year undergraduate and have been part of the “Great Game” – as observer and participant – ever since. The early 1970’s very a period of lively intellectual foment, heated discussions, and mental gymnastics, as the leading practitioners of the day attempted to rectify the new mobilist theory with existing concepts like geosynclines and morphotectonic belts. The new terrane paradigm arose out of this riotous, highly fertile period. In 1977, Davy Jones and colleagues named and outlined the first terrane, Wrangellia. Three years later, Peter Coney, Jones and Jim Monger, who had been separately and collectively wrestling with different parts of the Cordillera, published their seminal paper, Cordilleran Suspect Terranes (1980). In this paper, for the first time, we see the Cordillera carved up into these new entities, which are defined as internally coherent but mutually discrete; possibly all of vastly different origins. The terrane map changed our world and has formed the basis of our thought ever since. Of course, once established, every mental construct is subject to immediate and ongoing challenge. My colleagues and contemporaries have engaged in testing the terrane framework from its inception to the present day; and the current generation of students are continuing the work. A key theme of revision has been the recognition of linkages between terranes that significantly predate their accretion. In the original model, each terrane was considered an

The lithospheric mantle should be depleted in base- and precious-metals as these elements are transferred to the crust during partial melting. However, some melt-depleted mantle peridotites are enriched in these ore-forming elements. This may reflect re-fertilization of the mantle lithosphere and/or sequestering of these elements by residual mantle phase(s). Both processes remain poorly understood because of the low abundances of incompatible elements in peridotite and the nugget-like distribution of digestion-resistant mantle phases that pose analytical challenges for conventional geochemical methods. Herein we report new major and trace element concentrations for a suite of mantle peridotite and pyroxenite samples from the Late Permian to Middle Triassic Nahlin ophiolite (Cache Creek terrane, British Columbia, Canada) using Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICPMS) analysis of nanoparticulate powders and olivine. Compatible to moderately incompatible element concentrations suggest that Nahlin ophiolite peridotites represent residues after >= 20% melt extraction. Pyroxenite dykes and replacive dunite bands are folded and closely intercalated with residual harzburgite. These field relationships, coupled with the presence of intergranular base metal sulphide, clinopyroxene and Cr-spinel at the microscale, point to percolating melts that variably re-fertilized melt-depleted mantle peridotite. Radiogenic Pb (Pb-206/Pb-204 = 15.402-19.050; Pb-207/Pb-204 = 15.127-15.633; Pb-208/Pb-204 = 34.980-38.434; n = 45) and Os (Os-187/Os-188 0.1143-0.5745; n = 58) isotope compositions for a subset of melt-depleted peridotite samples further support metasomatic re-fertilization of these elements. Other ore-forming elements are also implicated in these metasomatic reactions because some melt-depleted peridotite samples are enriched relative to the primitive mantle, opposite to their expected behaviour during partial melting. New LA-ICPMS analysis of fresh olivine further demonstrates that a significant proportion of the highly incompatible element budget for the most melt-depleted rocks is either hosted by, and/or occurs as trapped inclusions within, the olivine-rich residues. Trapped phases from past melting and/or re-fertilization events are the preferred explanation for unradiogenic Pb isotope compositions and Paleozoic to Paleoproterozoic Re-depletion model ages, which predate the Nahlin ophiolite by over one billion years.

Late Triassic to Early Jurassic porphyry Cu mineralization is common in British Columbia, yet there are few age-equivalent porphyry occurrences in Yukon. This study presents new data for the enigmatic Carmacks Copper Cu-Au-Ag deposit in south-central Yukon, Canada, which is hosted in amphibolite facies metamorphic inliers within the Early Jurassic Granite Mountain batholith. Sulfide mineralization occurs mainly as net-textured bornite and chalcopyrite in leucosome, and as chalcopyrite ± pyrite blebs and disseminations in amphibolite and quartz-plagioclase-biotite schist. Several studies suggest that the Carmacks Copper deposit and the nearby Minto deposit are related to porphyry belts in British Columbia, but constraining the timing of alteration, mineralization, and metamorphism has been difficult. This study establishes a geologic and high-precision geochronologic framework for sulfide mineralization and its host rocks at the Carmacks Copper deposit, using Re-Os dating of molybdenite, and chemical abrasion-thermal ionization mass spectrometry (CA-TIMS) analysis of both whole zircon grains and laser-cut fragments of complexly zoned zircon grains. Our data indicate that the igneous protolith of the metamorphic inliers formed at 217.53 ± 0.16 Ma, followed by peak metamorphism at amphibolite facies at 205.82 ± 0.23 Ma, which occurred prior to Granite Mountain batholith emplacement but subsequent to Cu-Au-Ag mineralization of the protolith. An early phase of the Granite Mountain batholith was emplaced at 199.84 ± 0.14 Ma, followed by the main phase at 195 to 194 Ma. A second generation of metamorphic zircon in migmatite at 196.01 ± 0.12 Ma represents a partial melting event associated with Granite Mountain batholith emplacement. Two petrographically distinct populations of molybdenite are present in unstrained, net-textured copper sulfides. A sample dominated by strained molybdenite yielded an 187Re/187Os age of 212.5 ± 1.0 Ma, which represents the minimum mineralization age of the protolith. A sample dominated by euhedral grains yielded an 187Re/187Os age of 198.5 ± 0.9 Ma, constraining the maximum age of sulfide remobilization. These results indicate that primary mineralization is >212.5 Ma and potentially coeval with the ~217.5 Ma generation of Late Triassic magmatism. The mineralized protolith, best interpreted as the potassic alteration zone of a Late Triassic (~217–213 Ma) porphyry Cu-Au system, was metamorphosed to amphibolite facies at ~206 Ma, and subsequently migmatized during 200 to 194 Ma intrusion of the Granite Mountain batholith. The chalcopyrite-bornite-dominant assemblage in neosome precipitated from an immiscible Cu-Fe-S melt phase that partly consumed xenocrystic molybdenite and reprecipitated new molybdenite grains. The Carmacks Copper deposit and the related Minto deposit are remnants of a Late Triassic porphyry belt, where a significant fraction of the original metal endowment was likely lost through digestion of mineralized rocks by midcrustal magma in the Early Jurassic. These Yukon deposits are rare examples of metamorphosed porphyry Cu systems in the global geologic record, where rapid tectonic burial following mineralization was the principal factor in their preservation.

Ultramafic massifs from the Atlin (formerly Cache Creek) terrane, northern Canadian Cordillera, display an incomplete ophiolitic sequence as upper crustal rocks generally lie directly on mantle peridotites. Low-angle normal faults characterized by foliated serpentinites, rodingites and cataclasites mark the mantle-crust contact. Serpentinites from this contact at Union Mt. and Squanga Lake were investigated by microscopic petrography and Raman spectroscopy. Raman methods allow phase identification of different serpentine minerals: lizardite-crysotile (0-300°C; <1.0GPa) and antigorite (300-460°C; higher-P). Boron isotopic compositions of serpentine minerals are representative of the fluids that altered the ultramafic rocks. Previous work showed that serpentine hydrated by seawater has δ11B ranging from +40 to +10‰, while serpentine hydrated by slab-derived metamorphic fluids has δ11B ranging from +10 to -6‰ (shallow fluids) or -6 to -20‰ (deep fluids). At least three generations of serpentine minerals from the Atlin terrane were identified and analysed for δ11B: 1) early lizardite-chrysotile with δ11B values of +13 to -1‰; 2) a second generation of antigorite-chrysotile with δ11B values of -2 to - 13‰; and 3) late antigorite flakes with δ11B values of +4‰ to -2‰. Identification of several isotopically distinct serpentine generations support the hypothesis of: (1) early alteration of peridotites (early serpentine with +δ11B) by seawater-derived fluids, possibly associated with exhumation during oceanic core complex formation; (2) possible infiltration of aqueous fluids (2nd generation of serpentine with strongly negative δ11B) during obduction and imbrication of ophiolite nappes; (3) re-exhumation (late antigorite flakes with δ11B straddling 0‰), possibly reflecting late stages of obduction and thrust-stacking.

Serpentinization of ultramafic rocks drives geochemical exchange between the hydrosphere, biosphere and lithosphere in surficial environments and at depth. Precious metals are implicated in these hydration reactions because of the well-known association between peridotite-hosted Au and placer platinum-group mineral (PGM) deposits sourced from serpentinized ophiolite complexes at surface. However, the distribution of precious metals in mantle rocks and their mode of occurrence at the onset of serpentinization are not well understood because early-stage features are typically obliterated with progressive hydration. Herein we report electron probe microanalysis (EPMA) and laser ablation inductively coupled mass spectrometry (LA-ICPMS) spot and mapping results for a suite of base metal sulphide (pentlandite, pyrrhotite, chalcopyrite), native metal (Cu and Fe) and Ni-Fe alloy (awaruite) from variably serpentinized peridotite and pyroxenite (Late Permian to Early Triassic Nahlin ophiolite, Cache Creek terrane; Atlin, British Columbia, Canada). Pentlandite and pyrrhotite occur with magmatic clinopyroxene and Cr-spinel as ultrafine inclusions (few gm) and as coarser interstitial base metal sulphides (<= 200 mu m) that are enveloped by native metal (Cu and Fe) and Ni-Fe alloy. Because native Fe-and awaruite-bearing mineral assemblages require reduced and/or low-fS(2) conditions, we suggest that these replacement textures document destabilization of base metal sulphide phases during the conversion of olivine to serpentine and magnetite. Desulphurization reactions decoupled precious metals from relict pentlandite at the microscale, with Ag, Pd, Pt and Au partitioning from sulphides into both awaruite and native metal (Cu and Fe) at concentrations up to 100s of ppm. New high-resolution LA-ICPMS maps also point to clusters of ultrafine PGM, tellurides, bismuthides and metal alloys that were either remobilized within the serpentinized mesh and/or represent the completely desulphidized product of ultrafine, intergranular base metal sulphide. Other PGE (Os and Ir) are mostly hosted within relict pentlandite and/or pyrrhotite that were preserved during incomplete desulphurization, along with lesser microscale remobilization of these elements into ultrafine veins. New whole-rock PGE (nickel-sulphide fire-assay; NiS-FA) results demonstrate that precious metal remobilization was limited to the microscale during the earliest stages of serpentinization, which, in the case of Au, is consistent with its low solubility within such reduced, low fS(2) fluids. Precious metal mobility during the early stages of serpentinization may therefore depend on the stability of Ni-Fe alloy, native metal (Cu and Fe) and PGM. Progressive serpentinization, complete replacement of olivine and destabilization of the reduced, low-fS(2) mineral assemblage likely represents an important process to liberate, transport and concentrate precious metals within more oxidized and/or sulphur-bearing fluids in surficial and deep subduction environments. Crown Copyright (C) 2019 Published by Elsevier B.V. All rights reserved.

The Permo-Triassic Nahlin ophiolite is the largest and best-preserved ophiolite in the Canadian Cordillera of British Columbia and Yukon, Canada. The ophiolite is well-exposed along its ~ 150 km length with mantle segments divisible into the Hardluck and Menatatuline massifs. Both massifs comprise mostly depleted spinel harzburgite (< 2 wt% Al 2 O 3 and ~ 45 wt% MgO). Chondrite normalized REE abundances in clinopyroxene vary in (Gd/Yb) N from 0.2 to 1.1. Inversion modelling of clinopyroxene REE abundances requires 10–16% and 16–20% partial melting in the Hardluck and Menatatuline massifs, respectively. The two-pyroxene and Fe–Mg exchange temperatures in the mantle of the ophiolite also change systematically along strike with the degree of partial melt depletion. The temperatures recorded by REE and Ca-Mg exchange between coexisting pyroxenes require markedly higher peak temperatures and cooling rates for the Menatatuline massif (1250 °C, 0.1–0.01 °C/year) compared to the Hardluck massif (< 1100 °C, ~ 10 − 4 °C/year). The differences between these two contiguous massifs can be reconciled by their evolution as two separate segments along a ridge system having varying melt depletion, with contrasting cooling rates controlled by presence or absence of a crustal section above the mantle lithosphere, or by rapid exhumation along a detachment.