The juxtaposition of the composite Pearya terrane and the northern Laurentian margin at Ellesmere Island, Nunavut, Canada, has significant ramifications for the Paleozoic tectonic history of the circum-Arctic region. Published tectonic models rely upon interpretation of the subduction-related Kulutingwak Formation as an indicator of Ordovician and/or Silurian accretion (Trettin, 1998). New igneous and detrital zircon U-Pb and Lu-Hf isotopic data from 16 samples collected in the Yelverton Inlet–Kulutingwak Fiord region of northern Ellesmere Island suggest that the Kulutingwak Formation of Trettin (1998) contains structural blocks derived from both the Pearya terrane and Silurian strata associated with the ancestral Laurentian margin. Data from this study demonstrate a complex provenance history for rocks within the Petersen Bay, Kulutingwak Fiord, and Emma Fiord fault zones, with age probability peaks of ca. 470 Ma, 650 Ma, and 960–980 Ma that suggest affinity with the Pearya terrane, and age probability peaks of ca. 1800 Ma and 2700 Ma that indicate connections to the Laurentian margin. The combination of these signatures in Kulutingwak Formation rocks suggests that the Pearya terrane was proximal to the northern Laurentian margin by Late Ordovician time. Silurian and younger strike-slip displacement on the major fault zones resulted in the incorporation of blocks derived from the Pearya terrane basement and Silurian clastic rocks into the Kulutingwak Formation. Silurian displacement along these strike-slip faults, which are integral components of the Canadian Arctic transform system, is recorded by syndepositional deformation structures in the Danish River Formation and prevented the transition from soft to hard collision of the Pearya terrane. The two-stage model for the Pearya terrane—accretion followed by significant translation—provides a process for developing complex steep terrane boundaries with contentious displacement histories that are common in accretionary orogens.

Models for the tectonic evolution of the Arctic margin of North America suggest that large-scale (thousands of kilometers) displacement of terranes along the margin occurred in Paleozoic time (Colpron and Nelson, 2009; McClelland et al., 2023, and references therein). The concept is based on the recognition that some terranes in the northern Cordilleran orogen originated in the Iapetus realm that evolved between Baltica and Laurentia (e.g., Patrick and McClelland, 1995; Amato et al., 2009; Miller et al., 2011; Beranek et al., 2013a, Strauss et al., 2013). Direct evidence for Paleozoic terrane displacement includes the presence of the Pearya composite terrane outboard of deep-water passive margin sedimentary rocks on northern Ellesmere Island, Nunavut, Canada (Fig. 1). The Pearya terrane was originally interpreted as a northern continuation of Laurentian Precambrian basement (Schuchert, 1923; Thorsteinsson and Tozer, 1960), and some tectonic models still promulgate little to no displacement of the Pearya terrane (Hadlari et al., 2014; Dewing et al., 2019). However, Trettin (1987, 1998) convincingly argued that the Pearya terrane represents a displaced fragment emplaced on the Laurentian margin. The interpretation of significant terrane displacement allows for models that specifically correlate displaced Ordovician arc units in the Arctic–Alaska–Chukotka terrane of the Cordilleran orogen (e.g., Doonerak arc) with arc units of similar age in the Pearya terrane, and other arc complexes preserved in terranes of Svalbard and nappes of the Scandinavian Caledonides to the east (Fig. 1; Strauss et al., 2017, and references therein). In addition, the displacement model allows for involvement of terranes with similar Ordovician arc histories such as the Alexander terrane, which resides in the Insular composite terrane of the Cordilleran orogen (Fig. 1) but has origins in Baltica (Beranek et al., 2013a, 2013c; White et al., 2016). Although structures such as the Canadian Arctic Transform System (CATS; Fig. 1; McClelland et al., 2021, 2023) have been proposed to accommodate the juxtaposition and translation of Ordovician arc fragments along the northern Laurentian margin, the tectonic setting and timing of terrane displacement remain uncertain. As the only remaining fragment of the Ordovician arc system on the Laurentian Arctic margin (Trettin, 1998), the accretionary and displacement history of the Pearya terrane plays a critical role in resolving the uncertainty in translation models for northern Laurentia and evaluating the Paleozoic tectonic evolution of the circum-Arctic region.

Existing models for emplacement of the Pearya terrane on the northern Laurentian margin invoke orthogonal collision, strike-slip displacement, or some combination of the two in the Silurian (Bjornerud and Bradley, 1992; Klaper and Ohta, 1993; Trettin, 1987, 1998). Orthogonal collision is supported by the presence of Ordovician–Silurian volcanic and volcaniclastic rocks, marble, serpentinite-bearing conglomerate, and sandstone units assigned to the Clements Markham belt that separates the Pearya terrane in the north and Laurentian continental margin or Franklinian basin to the south (Fig. 1; Trettin, 1987, 1998). Collectively, the Clements Markham belt units have been interpreted as a subduction-related assemblage involved in a collisional event between the Pearya terrane and Laurentia (Klaper, 1992; Trettin, 1998). The presence of Ordovician arc rocks in the Pearya terrane and within the Kulutingwak Formation and other units of the Clements Markham belt suggest active convergence in the Ordovician. However, serpentinite-bearing units have been assigned to the Silurian Phillips Inlet Formation based on regional stratigraphic correlations (Trettin, 1998) that would indicate active subduction in the Silurian.

The subduction-related assemblage in the Clements Markham belt is overlain by a thick section of Silurian turbidites that has long been recognized as part of an overlap assemblage deposited in an elongate, margin-parallel basin (Trettin, 1987, 1998). The basin extends eastward to North Greenland, where Silurian turbidites mark input from the Caledonian orogen (Fig. 1; Surlyk and Hurst, 1984). Paleocurrent data from the basin record axial input from the Greenland Caledonides as well as the Pearya terrane and Franklinian basin sources (Trettin, 1994, 1998). Provenance studies based on detrital zircon age spectra are consistent with this interpretation (Anfinson et al., 2012; Hadlari et al., 2014; Beranek et al., 2015; Malone et al., 2019). Trettin (1998) concluded that accretion of the Pearya terrane occurred prior to overlap by the Danish River Formation in the Early Silurian (latest early Llandovery). However, soft sediment deformation structures present in Silurian turbidites are interpreted to record active subduction and accretion of the Pearya terrane, which challenges this post-accretion overlap hypothesis (Bjornerud and Bradley, 1992). Given the structural complexity, lack of stratigraphic continuity, and poor age control on the Clements Markham belt lithologies, their role in establishing an Ordovician versus Silurian age for the juxtaposition of the Pearya terrane and Laurentian Arctic margin remains ambiguous. For example, was accretion of the Pearya terrane accomplished by a two-stage process of Ordovician oblique convergence followed by Silurian strike-slip displacement, or by prolonged Ordovician–Silurian oblique convergence?

These regional associations highlight the difficulties in evaluating tectonic models in structurally complex regions where stratigraphic relationships and depositional ages are poorly established. Interpretation of this remote area in Ellesmere Island is complicated by a prolonged deformation history that includes brittle reactivation of Paleozoic faults in the Cenozoic (Powell and Schneider, 2022). Detrital zircon U-Pb-Hf isotopic signatures offer a decisive tool for discriminating among highly deformed packages present in the Yelverton Inlet–Kulutingwak Fiord region. This study presents new igneous and detrital zircon U-Pb and Lu-Hf isotopic results from 16 samples from northern Ellesmere Island to better understand the depositional age, provenance, and regional significance of the Kulutingwak Formation and related units. The samples were collected along a north–south transect adjacent to Yelverton Inlet and Kulutingwak Fiord (Fig. 2). Based on our new data, we propose that the Kulutingwak Formation, as defined by Trettin and Frisch (1987) and Trettin (1998), is a structurally complex unit that hosts a variety of displaced blocks from both the Pearya terrane and Silurian sedimentary units with provenance ties to the Arctic Laurentian margin. We also suggest that Ordovician rocks of the Clements Markham belt are related to the Pearya terrane and together represent an arc that accreted to the Laurentian margin during oblique subduction, whereas the Silurian turbidite sequence of the Danish River Formation was deposited in a strike-slip system that formed after the Ordovician arc collision.

Pearya Terrane

The Pearya terrane is composed of a series of fault-bounded crustal fragments imbricated against Paleozoic rocks of the Franklinian basin that mark the northern Laurentian margin (Fig. 1). Although its accretionary history is debated, Mesoproterozoic to late Paleozoic assemblages of the Pearya terrane are lithologically and geochronologically distinct from the adjacent Franklinian basin and the underlying Archean to Paleoproterozoic crystalline basement rocks of northern Laurentia (Trettin, 1987, 1998; Anfinson et al., 2012; Malone et al., 2014, 2017, 2019; Gilotti et al., 2018). Based on igneous geochemistry and zircon U-Pb ages, fragments of the Pearya terrane have been compared or correlated with other Arctic locales that record Tonian (ca. 960–980 Ma) arc magmatism, such as Svalbard and Arctic Alaska (Malone et al., 2017; McClelland et al., 2019), and Ediacaran (ca. 525–650 Ma) magmatism similar to that recognized in the Timanide orogen of northern Baltica (Estrada et al., 2018; Malone and McClelland, 2020; Majka et al., 2021) and the Seward Peninsula (Amato et al., 2009) in the Arctic Alaska–Chukotka terrane (Fig. 1).

Trettin (1987) divided the Pearya terrane into five successions. Succession I consists of Mesoproterozoic to Neoproterozoic orthogneiss and subordinate metasedimentary rocks of Tonian or older age (Malone et al., 2017; Estrada et al., 2018). Faults separate the Tonian orthogneiss-dominant Succession I from Neoproterozoic to Cambrian metasedimentary rocks of Succession II (Trettin, 1987). Succession II metasedimentary rocks, including phyllite, marble, diamictite, quartzite, and schist, lack formal stratigraphy and are instead classified in terms of “lithologic units” and “preliminary stratigraphic units of local extent” (Trettin, 1998). Malone et al. (2014) used detrital zircon U-Pb geochronology to identify three subgroups in Succession II. Many Succession II clastic rocks yield detrital zircon ages of between ca. 1000 Ma and 2000 Ma, with variable input from ca. 550–650 Ma and ca. 970 Ma Tonian sources (Malone et al., 2014; Estrada et al., 2018). Succession III consists of Middle to Upper Ordovician volcanic and volcaniclastic strata of the Maskell Inlet Complex and ultramafic to granitic plutons of the Lower–Middle Ordovician Thores Suite (Trettin, 1998). The Maskell Inlet Complex is composed of andesitic to basaltic flows, tuffs, and volcaniclastic and sedimentary rocks in fault contact with Succession II rocks. Thores Suite igneous rocks are interpreted to represent an Ordovician sub-arc plutonic complex, and zircon from plutonic rocks yields U-Pb ages of between 469 Ma and 481 Ma (Trettin, 1998; Estrada et al., 2018; Majka et al., 2021).

Successions II and III were locally deformed and metamorphosed during the M’Clintock orogeny, which is recorded by greenschist- to amphibolite-facies Barrovian metamorphism and local granitoid plutonism (Trettin, 1998). The M’Clintock orogeny has previously been interpreted as a collision between an Early to Middle Ordovician arc and the Pearya terrane basement, broadly analogous to the Taconic and Grampian arc-collisions in the Appalachian-Caledonian orogen (Trettin, 1998). However, Majka et al. (2021) reported that igneous zircon in Thores Suite plutonic rocks contained inherited Cambrian to Paleoproterozoic cores, which indicates that the Thores Suite Ordovician arc complex was built upon a crustal fragment, potentially with Tonian basement and assemblages such as metasedimentary rocks preserved in Succession II. The timing of the M’Clintock orogeny is loosely constrained by a post-tectonic ca. 463 Ma granitic intrusion (Trettin, 1987).

Metamorphic rocks of successions II and III are separated from overlying Ordovician units of Succession IV by an angular unconformity. Succession IV includes a basal conglomerate containing clasts derived from the Thores Suite, shallow-marine siliciclastic strata, volcanic rocks, and dolostone (Trettin, 1998). Succession IV volcaniclastic units yield ca. 450–500 Ma detrital zircon grains (Hadlari et al., 2014; Malone et al., 2019). Upper Ordovician to Silurian siliciclastic and carbonate rocks of Succession V unconformably overlie successions II and IV. Middle and Upper Ordovician rocks of the Cape Discovery and Taconite River formations in successions IV and V have detrital zircon U-Pb ages that are dominated by unimodal ca. 462 Ma and 453 Ma subgroups, respectively (Malone et al., 2019), and locally include ultramafic detritus (Trettin, 1998). Silurian turbidite-bearing strata of the Cranstone and Danish River formations and siliciclastic and carbonate rocks of the Lands Lokk and Marvin formations are considered to depositionally overlie the Cape Discovery and Taconite River formations (Trettin, 1998).

Franklinian Basin and Clements Markham Belt

The Franklinian basin developed along the northern margin of Laurentia during protracted Cryogenian–Ediacaran rifting associated with the final break-up of Rodinia (Dewing et al., 2004; Macdonald et al., 2010, 2023; Cox et al., 2015; Ernst et al., 2016; Faehnrich et al., 2023). This passive margin succession contains over 11 km of Neoproterozoic to Devonian siliciclastic and carbonate rocks and is underlain by the Archean to Paleoproterozoic Canadian shield (Frisch and Trettin, 1991; Dewing et al., 2008). Passive margin strata of the Franklinian basin have generally been divided into shallow-water shelf facies in the south and deep-water continental slope and rise facies to the north (Trettin, 1994; Dewing et al., 2008, 2019).

Lower Paleozoic rocks of the Franklinian basin are separated from the Pearya terrane by a complexly deformed and diverse assemblage of sedimentary and volcanic rocks that comprise the Clements Markham belt (Fig. 1). Southern exposures in the Clements Markham belt assigned by Trettin (1998) to the Grant Land and Yelverton formations are interpreted to represent distal Ediacaran–Cambrian rift and passive margin units tied to Laurentia (Faehnrich et al., 2023). A northern belt of Ordovician–Silurian volcanic and volcaniclastic units, including those of the Kulutingwak and Phillips Inlet formations (Trettin, 1998), has been variably assigned to the Pearya terrane or Laurentian origins (Bjornerud, 1991). Silurian clastic units of the Danish River and Lands Lokk formations extend from the Pearya terrane across the Clements Markham belt to the Laurentian margin (Fig. 1).

The Ordovician–Silurian volcanic and volcaniclastic units are exposed in 1–5-km-wide, fault-bounded, subplanar belts along major fault zones in the Clements Markham belt. From north to south, they are the Petersen Bay fault zone, Kulutingwak Fiord fault zone, and Emma Fiord fault zone (Fig. 2). The nomenclature for units exposed within and adjacent to these major fault systems varies in the literature (see Supplemental Material File S11 for a complete discussion) and culminates with the regional synthesis of Trettin (1998), which combined several isolated stratigraphic sections into the Ordovician Kulutingwak Formation and separated them from the serpentinite-bearing Silurian Phillips Inlet Formation. Trettin (1998) retained unique map-unit designations for other similar sections due to uncertainty in correlation (units Oks and OSv, Fig. 2). We use the map unit terminology of Trettin (1998) in this contribution but focus on refining the structural complexity and age relationships across the study area. At the original type locality of the Kulutingwak Formation along the Kulutingwak Fiord fault zone (map unit Ok, Fig. 2), Trettin (1998) described a 300–400-m-thick package of Ordovician sedimentary and volcanic rocks, including mafic to intermediate flows and tuffs, diverse siliciclastic units, serpentinite-bearing conglomerate and sandstone, and marble. The depositional age of the Kulutingwak Formation is based on an Ordovician (450 +4/–9 Ma) U-Pb age using multigrain thermal ionization mass spectrometry (TIMS) analysis of zircon from an isolated andesite-bearing succession and several conodont collections with Middle–Late Ordovician age assignments (Trettin, 1998). Trettin (1998) assigned a Silurian age to undated serpentinite-bearing units of the Phillips Inlet Formation based on regional correlation with the Fire Bay assemblage that, at the time, was thought to be Silurian. However, the Fire Bay assemblage is now interpreted to be Ordovician (Koch et al., 2022), which implies that the Kulutingwak and Phillips Inlet formations are age equivalent.

Amphibolite-facies metamorphic rocks exposed along the Petersen Bay fault zone are lithologically similar to subgreenschist-facies units of the Kulutingwak Formation. Trettin (1998) noted this similarity and included them in map-unit Oks (Fig. 2). The high-grade rocks, also referred to as the Peterson Bay assemblage, are juxtaposed with the Tonian basement of the Pearya terrane and deformed at amphibolite facies (Klaper and Ohta, 1993). Recent estimates from rare garnet–kyanite–staurolite–amphibole garbenschiefer in the Petersen Bay assemblage give peak Barrovian conditions of 650 °C at <0.9 GPa, and U-Pb monazite dating pins garnet growth at 385–397 Ma (Kośmińska et al., 2022).

Volcanic, volcaniclastic, and marble blocks in the Emma Fiord fault zone included in map unit OSv (Fig. 2) were tentatively assigned an Early Silurian (Llandovery) to Ordovician age based on a conodont collection, while the age of the volcanic rocks was unknown. Trettin (1998) broadly correlated volcanic units along the Emma Fiord fault zone with those at Fire Bay, ~130 km to the southwest (Fig. 1). Recent igneous and detrital zircon U-Pb data from volcanic and volcaniclastic rocks indicate that the Fire Bay assemblage is Early to Middle Ordovician in age (ca. 470 Ma; Koch et al., 2022).

Fault-bounded lenses of the Kulutingwak Formation (map units Ok and Oks, Fig. 2) and volcanic unit OSv are commonly juxtaposed with the Silurian Danish River and Lands Lokk formations, which are laterally extensive siliciclastic packages characterized by deep-water marine turbidite deposits derived from the north and northeast (Trettin, 1998). The Danish River Formation has historically been described as a homogeneous white mica-bearing lithic arenite that is laterally continuous for ~350 km from north to south (Trettin, 1998). The detrital zircon U-Pb signature of the Danish River Formation includes variable Neoproterozoic and Paleozoic ages depending on sample location (Anfinson et al., 2012; Hadlari et al., 2014; Beranek et al., 2015; Malone et al., 2019; Koch et al., 2022). The Silurian Lands Lokk Formation is restricted to the Clements Markham belt and forms a siliciclastic succession similar to the Danish River Formation, but locally includes volcanic components (Trettin, 1998). At one locality, the Lands Lokk Formation is described as conformably overlying the Danish River Formation (Trettin, 1998), but this relationship is uncertain given their similar lithology and complex history of nomenclature change (cf. Trettin and Frisch, 1987; Trettin, 1998).

Ellesmerian and Eurekan Deformation

Silurian and older sedimentary rocks of northern Ellesmere Island were deformed during the Late Devonian–Carboniferous Ellesmerian orogeny, which resulted in intense folding and local metamorphism within the Franklinian basin from northern Greenland to the western Canadian Arctic Islands (Klaper, 1992; Trettin, 1998; Piepjohn et al., 2008, 2013). Ellesmerian structures are characterized by southwest-directed thrust faults and local sinistral strike-slip displacement (Klaper, 1992; Piepjohn et al., 2013; Piepjohn and von Gosen, 2018). Deformed rocks of the Ellesmerian orogen are overlain by Carboniferous to Paleogene deposits of the Sverdrup basin, which covers an area of over 200,000 km2 across the Canadian Arctic Islands (Embry and Beauchamp, 2019).

Ellesmerian structures were reactivated and overprinted during Paleogene Eurekan deformation, which affected both the previously deformed assemblages of the Franklinian and Sverdrup basins, as well as the Pearya terrane (Piepjohn et al., 2008, 2013; Powell and Schneider, 2022). Eurekan deformation produced an intracontinental fold belt with margin-parallel strike-slip displacement in rocks of the Franklinian basin (Piepjohn et al., 2013; Piepjohn and von Gosen, 2018). All of the faults shown in Figure 2 have a strong brittle overprint, including late fractures, sliding surfaces, and cataclasite, which can be attributed to this Cenozoic displacement history (Powell and Schneider, 2022). These brittle faults accommodated appreciable late displacement that significantly modified the Paleozoic geometry (Piepjohn and von Gosen, 2018).

Sixteen samples were collected along a transect in the Yelverton Inlet and Kulutingwak Fiord region from the Petersen Bay fault zone to the Emma Fiord fault zone (Fig. 2), focusing on the Kulutingwak Formation and related units. Silurian clastic rocks of the Danish River and Lands Lokk formations in the Clements Markham belt were also collected, as well as units mapped in the Pearya terrane adjacent to the Petersen Bay fault zone. Location information can be found in File S2 (see footnote 1).

Fourteen samples were prepared for U-Pb analysis by laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) at the Arizona Laserchron Center, University of Arizona, Tucson, Arizona, USA. Twelve sedimentary rock samples and one igneous rock sample were crushed, pulverized, and separated by density on a Gemini shaker table and in heavy liquids following magnetic separation at the University of Iowa, Iowa City, Iowa, USA. Zircon from each clastic sample was isolated from a random aliquot of the heavy mineral separate by hand picking and removal of non-zircon grains. Approximately 500–1000 grains were mounted in 2.54 cm epoxy pucks with natural zircon reference material. Approximately 80 zircon crystals were hand-picked from the igneous rock sample on the basis of clarity and lack of inclusions, visible core-rim relationships, and fractures prior to co-mounting with natural zircon reference material in a 2.54 cm epoxy puck. The mounts were polished to expose the interior of the crystals and imaged by cathodoluminescence (CL) on a Hitachi S-3400N scanning electron microscope (SEM) at the University of Iowa MATFab Facility to reveal zoning within individual grains and to characterize spot locations. Detrital zircon for one sedimentary rock sample (04LB17) was separated, selected, and mounted in epoxy at Memorial University, Newfoundland, Canada, using the same techniques. CL imaging of the mount at Memorial University was completed using a JEOL JSM 7100F SEM. Representative CL images for each sample are provided in File S3 (see footnote 1).

Zircon U-Pb data were acquired using a Photon Machines Analyte G2 ArF 193 nm excimer laser with a Nu Plasma high-resolution (HR) multicollector mass spectrometer at the Arizona Laserchron Center. Analytical spots were 20 μm or 30 μm in diameter, depending on the size of grains or domains within grains. The mounts were cleaned with a 1% HCl-1% HNO3 solution prior to analysis, and each spot was ablated with three 40–50-μm-diameter cleaning shots prior to data collection to remove common Pb and other contaminants from the surface of the zircon mount (Pullen et al., 2018). Samples were run at a rate of either 600 or 300 analyses per hour (6 s or 12 s per analysis, respectively) according to the methods described in Sundell et al. (2021). Fractionation factors and U and Th concentrations in unknown grains were calibrated using natural zircon reference material FC (1099 Ma; Schmitz and Bowring, 2001) as a primary standard and SL (563 Ma; Gehrels et al., 2008) and R33 (419 Ma; Mattinson, 2010) as secondary reference materials.

Data were reduced with the MATLAB-based program AgeCalcML (github.com/kurtsundell/AgeCalcML). Uncertainties for each analysis (2σ) include (1) internal or measurement uncertainty and (2) external or systematic uncertainty added in quadrature. The results were filtered through evaluation of discordance and uncertainty using the following parameters to exclude analyses from provenance discussion: analyses with >10% discordance, >5% reverse discordance, or with 2σ age uncertainty errors >10%. 207Pb/206Pb ages were preferred for analyses older than 1200 Ma; 206Pb/238U ages were preferred for analyses younger than 1200 Ma. Probability density plots were generated using AgeCalcML and Isoplot 3.70 in Microsoft Excel (Ludwig, 2008). The Concordia age for sample 17-14 was calculated using Isoplot 3.70 (Ludwig, 2008). Laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) data from samples and reference materials are presented in File S4 (see footnote 1). Inter-sample variability is discussed qualitatively by comparison of the relative peaks in age distributions. Quantitative comparison of detrital zircon data is based on Kolmogorov–Smirnov D values (Vermeesch, 2018a; Saylor and Sundell, 2016) and presented as multidimensional scaling plots using the DZmds routine of Saylor et al. (2018). Evaluation of the maximum depositional age of each sample is based on maximum likelihood age (MLA) estimates (Vermeesch, 2021) calculated using IsolpotR (Vermeesch, 2018b).

Two additional samples of volcanic and volcaniclastic rocks were analyzed by secondary ion mass spectrometry (SIMS) at the NORDSIM facility, Stockholm, Sweden. Zircons were separated at Stockholm University using techniques similar to those described above. Handpicked zircon mounted in epoxy was polished to expose the grain interiors, and cathodoluminescence (CL) images were obtained using a Hitachi S4300 scanning electron microscope (SEM) at the Swedish Museum of Natural History in Stockholm. U-Th-Pb zircon analyses were conducted using a 15 µm spot size on a CAMECA IMS 1270 ion microprobe at the NORDSIM facility, following protocols outlined in Whitehouse et al. (1999) and Whitehouse and Kamber (2005). U-Pb ratios were calibrated based on the analysis of natural reference material 91500 (1065 Ma; Wiedenbeck et al., 1995), and the age interpretation is based on Concordia ages calculated using Isoplot 3.70 (Ludwig, 2008). Data from samples analyzed at the NORDSIM facility are presented in File S5 (see footnote 1).

Representative grains from age groups of interest were selected for Lu-Hf isotopic analysis on the Nu Instruments HR LA-ICP-MS at the Arizona Laserchron Center following the protocols of Gehrels and Pecha (2014). Locations of Lu-Hf isotopic analyses overlapped previously analyzed U-Pb spots in an effort to analyze the same age domain, and spot diameters were set at 40 μm. Calibration of Lu-Hf analyses used natural zircon standards FC, R33, SL, Mud Tank, and Plešovice (Schmitz and Bowring, 2001; Mattinson, 2010; Gehrels et al., 2008; Woodhead and Hergt, 2005; Sláma et al., 2008). Age-εHft plots for this study were generated using the Hf plotter routine of Sundell et al. (2019). All Hf data are presented in File S6 (see footnote 1).

Samples were collected along a transect extending from the southern boundary of the Pearya terrane to the Silurian rocks of the Danish River and Lands Lokk formations south of the Emma Fiord fault zone (Fig. 2). The samples are described, from north to south, in three groups associated with the major fault zones (Petersen Bay fault zone, Kulutingwak Fiord fault zone, and Emma Fiord fault zone; Fig. 2). Outcrop photographs and photomicrographs of selected samples are shown in Figures 3 and 4. Outcrop photos of the remaining samples are provided in Files S7 and S8 (see footnote 1). Additional descriptions of lithology, mineralogy, and zircon zoning for each sample are presented in File S2.

Petersen Bay Fault Zone

Samples collected in the Petersen Bay fault zone include three paragneisses (17-18, 17-20, and 17-80) assigned to the Mitchell Point belt of Succession I in the Pearya terrane and three schists (17-02, 17-16, and 17-55) and an orthogneiss (17-14) collected from the Petersen Bay assemblage or map unit Oks, the metamorphic equivalent of the Kulutingwak Formation (Fig. 2). Samples from Succession I are amphibolite-facies tectonites that were interpreted in the field to have sedimentary protoliths based on the presence of 10–20-cm-scale compositional layering (Fig. 3). Compositional layering is discontinuous at the outcrop scale and locally defines steeply plunging tight to isoclinal folds. The metasedimentary units are juxtaposed with Tonian granitic augen gneiss of the Pearya terrane, either by brittle faults that truncate layering or biotite-rich, layer-parallel shear zones. Samples of the Petersen Bay assemblage and Danish River Formation show a gradient in metamorphic grade that increases from south to north across a relatively narrow (10–15-km-wide) zone toward the boundary with the Pearya terrane (Klaper and Ohta, 1993).

Sample 17-18 was collected from a 10-cm-thick, quartz-rich layer within biotite–quartz–feldspar paragneiss with centimeter-scale layering (Figs. 3A and 3B). Approximately 15% of the U-Pb analyses were discordant, with 80% passing the combined discordance and uncertainty filters. The detrital zircon U-Pb results show a major ca. 500–512 Ma age peak, lesser ca. 460 Ma, 478 Ma, and 525 Ma age peaks, a broad Proterozoic subset with ca. 1010 Ma, 1150 Ma, and 1830 Ma age peaks, and a small Archean age distribution centered at ca. 2700 Ma (Fig. 5). The data define an MLA estimate of 459 ± 5 Ma. Sample 17-20 was collected from an outcrop of thinly interlayered (2–20 cm) quartzite and biotite-rich paragneiss. The metasedimentary unit contains 1–5-m-thick sheets of granitic orthogneiss and is bound by massive orthogneiss to the north. U-Pb results from sample 17-20 define a major ca. 965 Ma age peak, a minor ca. 1100 Ma age peak, continuous ages of between ca. 1200 Ma and 1870 Ma, and a few Neoarchean ages (Fig. 5). Sample 17-80 was collected from 2–5-cm-thick quartzofeldspathic layers within biotite gneiss containing deformed leucosome represented by abundant feldspar and quartz augen. The U-Pb results define a dominant ca. 961 Ma age peak, subsidiary ca. 1107 Ma age peak, and other ages between ca. 1400 Ma and 1500 Ma (Fig. 5). Samples 17-20 and 17-80 provide similar MLA estimates of 895 ± 10 Ma and 922 ± 8 Ma, respectively.

Petersen Bay assemblage sample 17-02 was collected from a 5-m-thick layer of mylonitic psammite within biotite-amphibole schist located between two prominent marble layers (Fig. 3C). Embayed quartz boundaries are evidence of grain-boundary migration recrystallization; quartz is commonly interlayered with calcite (Fig. 3D). Approximately 20% of the U-Pb analyses are discordant; the remainder define major ca. 502 Ma and 1830 Ma age peaks and minor subgroups of ca. 650 Ma and 1000 Ma grains. The youngest two grains with 206Pb/238U dates of 445 ± 7 Ma and 469 ± 8 Ma are outliers of the age group that defines the ca. 502 Ma age peak (Fig. 5). Exclusion of these two analyses shifts the calculated MLA from 452 ± 8 Ma to 476 ± 8 Ma. Orthogneiss 17-14 was collected from a 50-cm-thick felsic metaigneous sheet within amphibole-feldspar mafic schist. This sample is characterized by a spread in dates from ca. 880 Ma to 995 Ma that defines a Pb-loss trajectory and Proterozoic grains of between ca. 1000 Ma and 1700 Ma in age that represent inherited cores (Fig. 6). Mutually overlapping concordant analyses define a Concordia age of 970 ± 4 Ma (mean square weighted deviation [MSWD] = 1.3; Fig. 6) that is interpreted to be the igneous crystallization age. Sample 17-16 (Fig. 3E) is a weakly foliated, amphibolite-facies schist comprising the dominant lithology that hosts units sampled at locations 17-02 and 17-14. Poikiloblastic, coarse-grained amphibole is abundant and contains quartz and feldspar of varying size and shape (Fig. 3F). Major detrital zircon U-Pb age peaks are at ca. 956 Ma, 1820 Ma, and 2700 Ma. Small peaks represent distributed 206Pb/238U dates of between ca. 455 Ma and 500 Ma, and 650 Ma and 700 Ma (Fig. 5). The youngest single-grain 206Pb/238U dates are 423 ± 15 Ma and 428 ± 17 Ma. MLA estimates vary from 425 ± 11 Ma to 464 ± 8 Ma with and without the youngest two grains, respectively. Sample 17-55 was collected from a 60-cm-thick metasandstone layer in a steeply dipping sequence of interlayered sandstone and siltstone. The detrital U-Pb age spectrum is characterized by a single major ca. 947 Ma age peak with smaller ca. 480 Ma, 660 Ma, 1100 Ma, and 2700 Ma age groups (Fig. 5). The youngest grain has a 206Pb/238U date of 441 ± 7 Ma. Exclusion of this single analysis shifts the calculated MLA from 441 ± 6 Ma to 487 ± 3 Ma.

Sample 17-60 was collected from a sequence of 2–20-cm-thick, interbedded metasandstone and metasiltstone in the Silurian Danish River Formation ~4 km south of the Petersen Bay fault zone (Fig. 2). Finer grained layers include garnet–biotite–white mica ± amphibole that record metamorphism seen along the Petersen Bay fault zone. Detrital zircon U-Pb analyses define a major ca. 960 Ma age peak and minor ca. 505–700 Ma and 1000–2600 Ma subgroups (Fig. 5). The youngest three grains have 206Pb/238U dates of 501 ± 8 Ma, 539 ± 8 Ma, and 558 ± 7 Ma. Exclusion of the three youngest analyses shifts the calculated MLA from 515 ± 8 Ma to 596 ± 10 Ma.

Kulutingwak Fiord Fault Zone

Sample 04LB17 was collected from an outcrop of fine- to coarse-grained sandstone along the Kulutingwak Fiord fault zone ~30 km south of Mitchell Point (Fig. 2; map unit Ok). This section is in fault contact with a marble layer within the Kulutingwak Formation to the northeast. Detrital zircon U-Pb analyses define ca. 680 Ma and 965 Ma age peaks with additional signatures between 1000 Ma and 1700 Ma and a few analyses between 2500 Ma and 2900 Ma (Fig. 7). The youngest grain, with a 206Pb/238U date of 444 ± 7 Ma, is considerably offset from the next youngest subgroup of four grains ranging from 491 ± 10 Ma to 531 ± 6 Ma. Exclusion of the youngest analysis shifts the calculated MLA from 443 ± 7 Ma to 497 ± 12 Ma. Sample VP17-04b was collected from the Kulutingwak Formation ~0.5 km northeast of sample 04LB17 (Fig. 2). This altered metavolcanic rock represents “subcrop” along a frost-shattered ridge of mixed metacarbonate and metavolcanic rocks. A single U-Pb age population yielded a Concordia age of 500 ± 3 Ma (MSWD = 0.8; Fig. 6), which is interpreted as the crystallization age.

Sample 17-39 was collected from the Danish River Formation ~1 km south of the Kulutingwak Fiord fault zone between Yelverton and Kulutingwak inlets (Fig. 2). The clastic section consists of interlayered, fine-grained sandstone, siltstone, and shale and is in fault contact with limestone breccia and serpentinite conglomerate of the Kulutingwak Formation to the north. Analyses define major ca. 670 Ma and 960 Ma age peaks, and additional age subgroups include a few grains from ca. 465–500 Ma, 1000–1850 Ma, and 2600–2800 Ma (Fig. 7). The 206Pb/238U dates of the three youngest grains are 466 ± 7 Ma, 490 ± 8 Ma, and 497 ± 25 Ma. Calculated MLA estimates shift from 468 ± 7 Ma to 498 ± 6 Ma with exclusion of the youngest analysis.

Emma Fiord Fault Zone

Samples KF17-125, KF17-128, KF17-129, and VP17-05b were collected along a detailed transect through map unit OSv, which is exposed within the Emma Fiord fault zone between Kulutingwak Fiord and Yelverton Inlet (Fig. 2). Sample KF17-125 was collected from a 3-m-thick lens of medium-grained sandstone in a strongly deformed unit characterized by dismembered sandstone layers in a mudstone matrix (Figs. 4A and 4B). Detrital zircon analyses define ca. 445 Ma, 650 Ma, 980 Ma, and 1000–1880 Ma age peaks and a small 2600–3100 Ma subgroup (Fig. 7). The youngest three grains, at 445 ± 7 Ma, 445 ± 7 Ma, and 452 ± 9 Ma, clearly define the youngest peak, with an MLA estimate of 446 ± 4 Ma (Fig. 7). Sample KF17-128 is from a 30–70-cm-thick lens of lithic pebble-cobble conglomerate within black shale (Fig. 4C). The age distribution is dominated by a single ca. 465 Ma age peak (Fig. 7). The two youngest 206Pb/238U dates, at 423 ± 9 Ma and 424 ± 12 Ma, lie outside of the ca. 465 Ma age peak distribution. Exclusion of the youngest two analyses has a negligible effect on the calculated MLA, shifting it from 442 ± 3 Ma to 444 ± 3 Ma. Sample KF17-129 was collected from an 80-cm-thick layer of coarse-grained, volcaniclastic sandstone (Fig. 4E). U-Pb analyses define a dominant ca. 460 Ma age peak and a smaller ca. 438 Ma age peak (Fig. 7). The youngest two grains have 206Pb/238U dates of 427 ± 8 Ma and 432 ± 7 Ma. Similar to sample KF17-128, exclusion of the youngest two analyses has a negligible effect and shifts the calculated MLA from 434 ± 4 Ma to 437 ± 4 Ma. Sample VP17-05b was collected near the southern limit of the Emma Fiord fault zone (Fig. 2) from a boudin of deformed volcaniclastic rock in shale. Zircon analyses define a Concordia age of 470 ± 3 Ma (MSWD = 1.4), which is interpreted as the crystallization age of the volcanic source material (Fig. 6).

Sample KF17-217 was collected ~5 km south of the Emma Fiord fault zone, near the head of Kulutingwak Inlet, from the Silurian Lands Lokk Formation (Fig. 2). The zircon yield a U-Pb age spread of ca. 440–500 Ma, with major peaks at ca. 650 Ma and 956 Ma (Fig. 7). The three youngest grains have 206Pb/238U dates of 438 ± 6 Ma, 440 ± 5 Ma, and 442 ± 5 Ma.

Lu-Hf Results

Cambrian–Ordovician (ca. 450–500 Ma) detrital zircon in the Kulutingwak Formation and correlated units mainly yielded superchondritic to subchondritic εHf(t) (where εHf(t) = 176 Hf/177Hf) values of −5 to +10 (Fig. 5), with the exception of a grain with a value of +13.3 from sample KF17-129 (Fig. 7). A notable exception is the juvenile Silurian and Ordovician signature from sample 17-55, which exclusively records εHf(t) values of between +5.6 and +10.4 (Fig. 5). Neoproterozoic (515–685 Ma) grains from samples 04LB17 and KF17-125 give εHf(t) values of between −6.8 and +2.8. Tonian grains with ages of between ca. 960 Ma and 980 Ma produced εHf(t) values of between −5 and +5, with the exception of a few grains from samples 17-80 and 17-20 at −10.2 and a grain from sample 17-14 at +9.2. Paleoproterozoic zircon is significantly more evolved, with εHf(t) values of between −13.9 and +7.3 for grains with ages of ca. 1860 Ma, and values of between −10.2 and +5.8 for grains with ages of ca. 2700–2900 Ma.

Paleozoic grains from the Danish River and Lands Lokk formations have Silurian and Ordovician εHf(t) values that typically range between −10 and +5. The samples analyzed also have a well-defined ca. 960–980 Ma cluster of analyses with values of between −5 and +5, with a few grains having higher values of up to +10. Stenian zircon (ca. 1050 Ma) yielded εHf(t) values of between 0 and +10.9. Mesoproterozoic and older grains typically record εHf(t) values of between −5 and +5 (Figs. 5 and 7).

Past field observations focused on the Kulutingwak Formation and related units in the Clements Markham belt, from initial mapping efforts (Trettin and Frisch, 1987) to more detailed studies (Bjornerud, 1989, 1991; Klaper, 1992; Klaper and Ohta, 1993), consistently characterized Kulutingwak Formation units as structural complexes as opposed to the coherent stratigraphic packages of Trettin (1998). The lithologic variation, contrasting degrees of deformation, and varied provenance described here for the Kulutingwak Formation and related units are consistent with the earlier observations. We discuss the seminal work of Trettin (1998) in the context of our field correlations and new igneous and detrital zircon data from the Petersen Bay fault zone, Kulutingwak Fiord fault zone, and Emma Fiord fault zone. We distinguish between detrital zircon signatures found in the Kulutingwak Formation proper (Type A) and those that are suspected structural blocks of surrounding units (Type B). This discrimination is made on the basis of U-Pb signatures found in the samples taken from surrounding units, including the Pearya terrane, the Ordovician Fire Bay Assemblage, and Silurian Danish River and Lands Lokk formations, and previously published zircon U-Pb data (Anfinson et al., 2012; Malone et al., 2014, 2017, 2019; Beranek et al., 2013b; Hadlari et al., 2014; Estrada et al., 2018; Koch et al., 2022).

Pearya Terrane

Samples 17-18, 17-20, and 17-80 were collected from the mapped extent of Succession I of the Pearya terrane, which is characterized by minor metasedimentary units intruded by widespread Tonian orthogneiss (Trettin, 1998). There are no reliable detrital zircon results from Succession I metasedimentary units, but the expectation is that the units should not contain ca. 965 Ma zircon. However, paragneiss samples 17-20 and 17-80 are dominated by ca. 965 Ma detrital peaks in a signature that is characteristic of Group B in Succession II as defined by Malone et al. (2014). These two samples are accordingly interpreted as metamorphosed Succession II units that are structurally intermixed with the adjacent orthogneiss of Pearya terrane Succession I in the Petersen Bay fault zone. In contrast, sample 17-18 has a different signature that is discussed below in the context of the Kulutingwak Formation.

Kulutingwak Formation (Type A)

The most striking detrital signatures from the Petersen Bay fault zone are found in samples 17-18, 17-02, and 17-16. Sample 17-18 lies within units mapped as Succession I, whereas samples 17-02 and 17-16 lie within the Petersen Bay assemblage just south of the mapped Pearya terrane boundary (Fig. 2). Sample 17-02 is from a fault block, and sample 17-16 represents the dominant lithology of the Petersen Bay assemblage. The detrital signature of these samples is nearly identical, with Mid–Late Cambrian (Miaolingian) ca. 502–508 Ma age peaks, as well as ca. 1800 Ma and 2700 Ma age peaks. On the basis of similarity in detrital zircon signatures, sample 17-18 is interpreted as a sliver of the Petersen Bay assemblage that is structurally intercalated with the basement gneiss of the Pearya terrane. The signature defined by the three samples is inferred to be the best record of the Peterson Bay assemblage, and by correlation (Trettin, 1998), the Kulutingwak Formation. Assuming the youngest grains are outliers due to Pb-loss, the abundant 445–470 Ma grains and MLA estimates from 459 ± 5 Ma to 476 ± 8 Ma point to an Ordovician protolith age for the Petersen Bay assemblage and Kulutingwak Formation.

Cambrian ages are not common or well documented in the Canadian Arctic—the Kulutingwak Formation signature stands out from both the Pearya terrane and the Franklinian margin. Exceptions within the Pearya terrane include a welded tuff interlayered with Succession II strata to the north of Milne Fiord, which yielded a multigrain TIMS zircon U-Pb age of 503 +8/–2 Ma (Trettin et al., 1987), the Ward Hunt pluton with a concordia age of 542 ± 2 Ma in the northern Pearya terrane (Malone and McClelland, 2020), and potentially Succession III volcanic and igneous rocks as suggested by Cambrian detrital zircon from Succession IV and V clastic rocks attributed to Succession III (Malone et al., 2019). Sources within the Clements Markham belt include a felsic volcanic clast dated from a conglomerate in the Ordovician Fire Bay assemblage that yielded a zircon U-Pb age of 498 ± 6 Ma (Koch et al., 2022), as well as sample VP17-04b from this study, an altered metavolcanic rock collected from the Kulutingwak Fiord fault zone that gave a zircon Concordia age of 500 ± 3 Ma. The prominent ca. 500 Ma signature could indicate enigmatic sourcing from along-strike variations in the Pearya terrane basement ages, earlier phases of magmatism recorded in the Ordovician arc basement, or extension-related magmatism along the Laurentian passive margin (Faehnrich et al., 2023). Notably, the Paleoproterozoic and Neoarchean signatures present in samples 17-18 and 17-02, namely the prominent ca. 1800 Ma and 2700 Ma age peaks, do not resemble Pearya terrane-derived strata of Succession II (Malone et al., 2014). This signature is found, however, in Neoproterozoic to Cambrian sedimentary rocks of the Franklinian margin (Kirkland et al., 2009; Anfinson et al., 2012; Beranek et al., 2013b; Dewing et al., 2019; Faehnrich et al., 2023).

Amphibole-rich schist, including garbenschiefer (Kośmińska et al., 2022), is the dominant lithology within the Petersen Bay assemblage along the Petersen Bay fault zone. The detrital zircon U-Pb age signature of this unit defined by sample 17-16 appears to combine the signature of the Pearya terrane with the unique age spectra found in samples 17-02 and 17-18 (Fig. 8). The ca. 956 Ma zircons were likely sourced from the Pearya terrane (Malone et al., 2017). Zircons with ages between ca. 600 Ma and 700 Ma may reflect derivation from Succession II and younger strata of the Pearya terrane (Malone et al., 2014; Estrada et al., 2018). Finally, a few grains of between ca. 456 Ma and 514 Ma may also be tied to Ordovician arc rocks of successions III and IV of the Pearya terrane (Trettin, 1998; Malone et al., 2019) or volcaniclastic rocks of the Fire Bay assemblage in the Clements Markham belt (Koch et al., 2022). The prominent 1800 Ma and 2700 Ma age peaks more closely resemble derivation from the Franklinian margin, as was proposed for samples 17-02 and 17-08. Thus, the Kulutingwak Formation (unit Ok) and correlatives in the Petersen Bay assemblage (unit Oks) are interpreted to record input from both the Pearya terrane and Franklinian margin.

Structural Blocks Assigned to the Kulutingwak Formation (Type B)

Samples described below illustrate a key part of the difficulty in understanding the origins and evolution of the Kulutingwak Formation, i.e., the displacement of structural blocks (Type B) due to a continued history of Paleozoic–Cenozoic translation. The most striking example is the highly recrystallized felsic metaigneous rock of sample 17-14 collected within the Peterson Bay assemblage (Fig. 2). The sample yielded a Concordia age of 970 ± 4 Ma and εHf(t) values of between −5 and +5 that are consistent with the Tonian age and εHf(t) signature of the Pearya terrane (samples 17-20 and 17-80, this study; Malone et al., 2017). Thus, sample 17-14 is interpreted as a block of Succession I of the Pearya terrane that is structurally intercalated with the Petersen Bay assemblage in the Petersen Bay fault zone.

Similarly, sample 17-55 was collected from the Petersen Bay assemblage (Fig. 2B) but is more lithologically similar to the Danish River Formation than the Kulutingwak Formation (Trettin, 1998), although it contains rare white mica (<1%) and up to 30% biotite. The detrital zircon U-Pb signature of sample 17-55 strongly resembles that of sample 17-60, which was collected from the Danish River Formation less than 5 km south of the Petersen Bay fault zone. A major detrital zircon ca. 947 Ma age peak demonstrates ties to Pearya terrane Succession I (Malone et al., 2017), whereas other Neoproterozoic to Archean age components are similar to those of Succession II (Malone et al., 2014). Smaller ca. 470–490 Ma and 660 Ma age subgroups are consistent with the Cambrian–Ordovician arc signature of the Pearya terrane as well. The slight offset of the major Tonian age peak between samples 17-55 and 17-60 (947 Ma versus 960 Ma, respectively) is interpreted to reflect Pb loss, which is consistent with the presence of metamorphic rims, irregular zoning, and a high degree of discordance in the detrital grains analyzed (220/600 analyses fail discordance filters). Although the detrital signature of sample 17-55 does share some characteristics with other samples of the Petersen Bay assemblage, it lacks important markers like the 1800 Ma and 2700 Ma age peaks that distinguish samples 17-02, 17-18, and 17-16 (Fig. 8). Based on comparison with sample 17-60, sample 17-55 is interpreted as a fault block of the Danish River Formation that is structurally intercalated with blocks of the Kulutingwak Formation and Pearya terrane basement rocks in the Petersen Bay fault zone.

Sample 04LB17, a sandstone in fault contact with deformed Kulutingwak Formation marble in the Kulutingwak Fiord fault zone, was collected from the fault zone where the Kulutingwak Formation of Trettin (1998) was originally described. However, the detrital zircon U-Pb signature of this sample, with age peaks of ca. 680 Ma and 965 Ma, is more similar to samples of the Silurian Danish River and Lands Lokk formations (17-39 and KF17-217, this study) than those of the Kulutingwak Formation described above (Fig. 8). Some of this discrepancy could be accounted for by sample size (04LB17, n = 155 versus 17-18, n = 457), but the large abundance of ca. 680 Ma grains compared to samples in the Petersen Bay fault zone may indicate the presence of a different source. Due to the structural setting of this sample and the similarity of the detrital zircon signature with nearby samples of the overlying Silurian sedimentary units, sample 04LB17 may be from a structural block of the Danish River or Lands Lokk formations that is displaced within the Kulutingwak Fiord fault zone.

In the Emma Fiord fault zone, sample KF17-125 was collected from an outcrop of immature lithic arenite and mudstone, a lithology that is distinctly different from the Ordovician volcanic and volcaniclastic samples of the Kulutingwak Formation. The detrital signature of sample KF17-125 is more varied, with major age peaks at ca. 656 Ma and 979 Ma and additional peaks at 446 Ma and 1665 Ma (Fig. 7). This sample is more similar to samples 17-39, 04LB17, and KF17-217, the latter of which is from the Silurian Lands Lokk Formation, and was collected ~5 km to the southwest of KF17-125. This sample also has similarities to detrital signatures found in the Danish River and Lands Lokk formations along the Emma Fiord fault zone near Fire Bay (Koch et al., 2022). U-Pb detrital zircon signatures suggest that sample KF17-125 may be a displaced block of the Silurian Lands Lokk or Danish River formations.

Structural Blocks Assigned to the Fire Bay Assemblage

At the Emma Fiord fault zone, a different but decidedly Ordovician detrital zircon age signature is recorded in unit OSv (Fig. 7). Volcaniclastic samples KF17-128 and KF17-129 are characterized by unimodal U-Pb age peaks at ca. 465 Ma and 456 Ma, respectively, but grains in both samples have likely experienced significant Pb loss. This is indicated by high-U (CL-dark) grains and a high-percentage of discordant analyses (n = 123/289 and n = 70/149, respectively). Thus, these Ordovician peaks correlate well with either arc-related rocks of the Pearya terrane, such as the Cape Discovery and Taconite River formations (Malone et al., 2019; Majka et al., 2021), or the Fire Bay assemblage in the Clements Markham belt (Fig. 8), which are defined by zircon U-Pb ages of ca. 465–480 Ma (Koch et al., 2022). This signature is also in agreement with the zircon Concordia age of 470 ± 3 Ma (Fig. 6) from sample VP17-05b, which was collected a few kilometers to the west. Samples KF17-128 and KF17-129 record zircon εHf(t) values of −5 to +10 that overlap with those of the Fire Bay assemblage (Fig. 7). Significantly, these samples within the Emma Fiord fault zone are ~130 km along strike from the type area of Fire Bay assemblage, which is locally deformed by the same fault system (Fig. 1; Koch et al., 2022). Juvenile εHf(t) signatures and the mafic geochemical compositions of Fire Bay assemblage rocks indicate origins in an intraoceanic arc setting, potentially between the Pearya terrane and the Doonerak arc assemblage of northern Alaska (Fig. 1; Strauss et al., 2017). The U-Pb-Hf signatures and structural positions of samples KF17-128 and KF17-129 suggest correlation with the Fire Bay assemblage rather than the Kulutingwak Formation.

Tectonic Setting of the Kulutingwak Formation

The Kulutingwak Formation of Trettin (1998) contains a specific assemblage of lithologies, including carbonates, volcanic flows and tuffs, shallow intrusions, metasedimentary rocks, and notably, both massive and sedimentary serpentinite that have chaotic and/or obscured structural contacts (Bjornerud, 1991). This diversity led to the interpretation that the Kulutingwak lithologies were deposited in an arc-related setting or subduction-zone complex (Klaper, 1992), with emphasis placed on the occurrence of lenticular bodies of serpentinite. In the Kulutingwak Fiord fault zone, sedimentary units composed of serpentinite-bearing conglomerate and sandstone (Trettin, 1998; Bjornerud and Bradley, 1992) are common, but Trettin (1998) assigned units with serpentinite to the Silurian Phillips Inlet Formation. The original Silurian age designation was based on regional correlation with the Fire Bay assemblage, which has been revised to Ordovician to reflect new age results from the Fire Bay volcanic and volcaniclastic succession (Koch et al., 2022). Our attempt to date a serpentinite-bearing conglomerate in the Phillips Inlet Formation unfortunately yielded no zircon. Nevertheless, we suggest that serpentinite-bearing lithologies are best interpreted as coeval with arc-related volcanic and volcaniclastic rocks within the Kulutingwak Formation and, therefore, are most likely Ordovician or older. This interpretation is consistent with the occurrence of ultramafic detritus in Ordovician units of the Pearya terrane (e.g., Taconite River Formation) and the Fire Bay assemblage (Trettin, 1998).

The volcanic affinity of the Kulutingwak and serpentinite-bearing Phillips Inlet formations suggests formation in a forearc, accretionary complex, and/or peripheral foredeep setting in the Ordovician, coeval with arc magmatism recorded in the Fire Bay assemblage and Pearya terrane (Fig. 9A). Our new detrital zircon U-Pb data from the Kulutingwak Formation indicate that this syn-collisional sedimentary system received detritus from both the Pearya terrane and Franklinian margin, as well as intervening Cambrian volcanic rocks. The Ordovician arc represented by the Pearya terrane and Fire Bay assemblage was likely impinging on the Laurentian margin by the Late Ordovician (Fig. 9A) rather than the Silurian as proposed in earlier tectonic models (Trettin et al., 1987; Klaper, 1992; Bjornerud and Bradley, 1992). The specific setting of the serpentinite-bearing units is unknown but is interpreted to perhaps be analogous to Mesozoic examples in California (Wakabayashi, 2017) or the serpentinite seamounts in the Marianas (Fryer, 2012); they may have been sourced from settings within the forearc of the Pearya terrane–Fire Bay arc system in the upper plate or extended Laurentian margin in the lower plate.

Identification of fault-bounded blocks (Type B) of Silurian clastic rocks (samples 17-55, 04LB17, and KF17-125) within the mapped extent of the Kulutingwak Formation and other Ordovician volcanic rocks (OSv) is significant because it shows that structural disruption of these Ordovician arc and subduction-related rocks occurred during or after the Silurian (Fig. 9B). Syn-depositional deformation structures in the Danish River Formation demonstrate that the tectonic boundary between the Pearya terrane and the Franklinian margin was still active during the Silurian (Bjornerud, 1991; Bjornerud and Bradley, 1992). These structures include abundant soft-sediment deformation, clastic dikes, and syn-sedimentary reverse faults with entrained bedding and small-scale folds that were interpreted to indicate that the Danish River Formation was deposited in an accretionary prism involved in Silurian arc-continent collision (Bjornerud and Bradley, 1992). However, the revised age of the Kulutingwak Formation suggests that all subduction-related rocks, including those from the Fire Bay assemblage and the Pearya terrane, are Ordovician (Koch et al., 2022; Malone et al., 2019). We suggest that these syn-sedimentary deformation features record post-collision deformation in a Silurian overlap sequence rather than the initial Pearya terrane collision (Fig. 9B).

Regional Significance

Our observations are consistent with the model of Bjornerud and Bradley (1992), where the distal Laurentian margin is the down-going plate, but suggest that collision began in the Ordovician rather than Silurian (Fig. 9A). The age of shear zones within the composite Pearya terrane suggests that sinistral deformation was ongoing in the Late Ordovician (McClelland et al., 2012). This intra-terrane deformation is consistent with oblique collision of the Ordovician arc represented by the Pearya terrane–Fire Bay arc with the Franklinian margin. The lack of observed shortening, crustal thickening, and metamorphism associated with this Late Ordovician event suggests a soft collision of the Pearya terrane with the Laurentian margin (van Staal and Zagorevski, 2020). We propose that Ordovician subduction ended during the oblique collision of the Pearya terrane with the distal Laurentian margin and ultimately resulted in conversion to a strike-slip boundary along the contact between the Pearya terrane and Laurentia (Fig. 9B).

Late Ordovician and younger strike-slip motion on major structures within the Pearya terrane and northernmost Franklinian basin is well-documented (Trettin, 1987, 1991; Piepjohn and von Gosen, 2018; McClelland et al., 2012). Thus, following Ordovician collision, we suggest that strike-slip displacement continued along the Laurentian margin (Fig. 9B), preventing the transition from a soft to hard collision phase (van Staal and Zagorevski, 2020). In this model, the overlapping Silurian Danish River Formation was deposited in a strike-slip or oblique retroarc foreland basin, which is consistent with the observed syndepositional deformation and dewatering structures in the Silurian clastic units (Bjornerud and Bradley, 1992). Deformation, disruption, and imbrication of Ordovician arc rocks of the Fire Bay assemblage and subduction-related rocks in the Kulutingwak Formation occurred in this Silurian strike-slip setting as well. Translation also offers a viable mechanism for emplacement of structural blocks of the Pearya terrane and Silurian sedimentary rocks into the Kulutingwak Formation. These observations are not only consistent with models depicting the northern Laurentian margin as a transcurrent boundary that accommodated sinistral terrane translation outboard of the Franklinian basin (McClelland et al., 2021, 2023), but are also supported by 40Ar/39Ar white mica ages from the Emma Fiord fault zone and Kulutingwak Fiord fault zone that record a prolonged history of deformation beginning in the Late Ordovician and continuing through the Cenozoic (Powell and Schneider, 2022). Thus, arc rocks preserved in the Pearya terrane, Fire Bay assemblage, and the Kulutingwak Formation represent an Ordovician arc system that collapsed against the northern Laurentian continental margin during Silurian and younger translation.

Amphibolite-facies equivalents of the Kulutingwak Formation found along the Petersen Bay fault zone were metamorphosed in the Early Devonian (ca. 385–397 Ma; Kośmińska et al., 2022). Local metamorphism of Kulutingwak Formation-equivalent schists in the Petersen Bay fault zone is coeval with widespread Ellesmerian deformation, which resulted in southwest-directed thrust faults, sinistral strike-slip displacement, intense folding, and regional metamorphism along the length of the Arctic margin (Piepjohn et al., 2013; Klaper and Ohta, 1993). In combination with the detrital zircon data from this study, this Devonian metamorphic event may offer a more tenable period of closure of the Silurian strike-slip basin system and final accretion of the Pearya terrane.

Implications for Crustal Structure

Our model envisions that the Clements Markham belt is largely underlain by Ordovician arc, forearc, and accretionary prism components represented by the Fire Bay and Kulutingwak formations, which are disrupted and juxtaposed by steep faults. This hypothesis raises the question of crustal structure at depth. Regional cross-sections (Piepjohn et al., 2013; Piepjohn and von Gosen, 2018) show the strike-slip faults rooting into a regional subhorizontal detachment associated with Ellesmerian and Eurekan shortening, where the Clements Markham belt and Pearya terrane now lie structurally above Laurentian crust. However, the mapped strike-slip faults spatially coincide with lateral variations in crustal composition defined by available geophysical data, at least to mid-crustal depths (Fig. 9C; Stephenson et al., 2018). We suggest that this relationship extends to lower crustal levels and infer that the juxtaposition of accreted material and Laurentian crust on steep faults within the Clements Markham belt is consistent with models of transpressional displacement along lithosphere-scale strike-slip faults during the Devonian–Carboniferous Ellesmerian orogeny. Additional geophysical experiments are required to evaluate the lower crustal and lithospheric relationships and advance models for the Paleozoic deformation history of northern Laurentia.

New igneous and detrital zircon U-Pb and Lu-Hf data from this study characterize the lithologies present in the Yelverton Inlet–Kulutingwak Fiord area of Ellesmere Island, a key location where relationships formed during the accretion of the Pearya terrane to the Franklinian margin are preserved and used to construct a stratigraphic framework for the region. Our data suggest that the Kulutingwak Formation of Trettin (1998) is an arc-related assemblage that records the proximity of the Pearya terrane to the Laurentian margin in the Late Ordovician. The data also demonstrate disruption of these arc-related rocks in Silurian and younger time via the inclusion of structural blocks of Silurian strata and Pearya terrane basement lithologies. Syn-depositional deformation structures in the Silurian Danish River Formation are interpreted to reflect strike-slip displacement of the Ordovician arc-related lithologies of the Kulutingwak Formation.

A two-stage model for accretion is proposed for the Pearya terrane. Ordovician rocks in the Clements Markham belt record the approach and soft collision of the Pearya terrane and related Ordovician arc complexes (Fire Bay assemblage) with the Laurentian margin through oblique subduction of the intervening oceanic lithosphere. Translation of the Pearya terrane continued in the Silurian as turbidites of the Danish River Formation were deposited in an active strike-slip basin. Devonian amphibolite-facies metamorphism of the Kulutingwak equivalents along the Petersen Bay fault may in turn record final closure of the Silurian basin that separated the Pearya terrane and the Laurentian margin, coeval with the onset of Ellesmerian deformation.

The scenario presented here for the Pearya terrane is consistent with the large-scale translation of terranes through the circum-Arctic region. Following accretion and dismemberment of the original Ordovician arc system, subduction and terrane displacement likely migrated outboard of the Pearya terrane to accommodate continued displacement and transfer of other terranes such as the Alexander terrane (White et al., 2016) to the Cordilleran orogen. Displacement on strike-slip faults after subduction-related basin closure and arc collision is expected to significantly modify the original accretionary boundary through the overprinting of older structures and redistribution of the various crustal fragments. Younger modification, like that observed here in the Canadian Arctic, produces complex boundaries with cryptic histories similar to many examples preserved in the Cordilleran orogen that remain contentious structures with uncertain displacement histories (Busby et al., 2023). Relationships documented for the Ordovician Kulutingwak and Silurian Danish River formations provide a tangible example of how complex accretion-translation terrane histories can be established for this type of boundary.

1Supplemental Material. File S1: Discussion of unit nomenclature in the Kulutingwak Fiord region. File S2: Sample locations, units, and extended descriptions. File S3: Representative cathodoluminescence images of zircon analyzed. File S4: Table of U-Pb isotopic data from the Arizona LaserChron Center. File S5: Table of secondary ion mass spectrometry U-Pb zircon geochronology. File S6: Table of Lu-Hf zircon geochemistry. File S7: Additional outcrop photographs of samples. File S8: Additional outcrop photographs of samples. Please visit https://doi.org/10.1130/GEOS.S.25505017 to access the supplemental material, and contact editing@geosociety.org with any questions.
Science Editor: Andrea Hampel
Associate Editor: Todd LaMaskin

Fieldwork on northern Ellesmere Island, Nunavut, Canada, was funded by the Bundesanstalt für Geowissenschaften und Rohstoffe (BGR) and the Geological Survey of Canada. We thank Karsten Piepjohn at the BGR for invitations to participate in this expedition (CASE19). Additional funding was provided by National Science Foundation (NSF) grants EAR-1650152 awarded to J.V. Strauss and EAR-1650022 awarded to J.A. Gilotti and co-principal investigator W.C. McClelland, Natural Sciences and Engineering Research Council of Canada Discovery Grants to L.P. Beranek (RGPIN 2014-06327 and 2019-05647), and Swedish Research Council Grant 1321731 to V. Pease. M.M. Koch received support from the Cook Scholarship fund from the Department of Earth and Environmental Sciences, University of Iowa. The Arizona Laserchron Facility is supported by NSF grant EAR-1649254. Reviews from Marcia Bjornerud, Jeff Amato, and Associate Editor Todd LaMaskin improved the manuscript.

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