Detrital zircon ages from upper Paleozoic–Triassic clastic strata on St. Lawrence Island, Alaska: An enigmatic component of the Arctic Alaska–Chukotka microplate

The Arctic Ocean region is underlain by Cretaceous and younger oceanic crust (e.g., Nikishin et al., 2021) that is surrounded by ancient cratonal landmasses of Laurentia, Baltica, and Siberia, as well as numerous smaller, tectonically juxtaposed continental fragments for which the origin and pre-breakup configuration are not well constrained (Fig. 1A). This is particularly true of the Arctic margins of North America and eastern Eurasia that surround the rim of the Amerasia Basin. The geologic history of continental fragments and associated continental shelves surrounding the Amerasia Basin (i.e., Arctic Canada, Arctic Alaska, Chukotka, and islands of the East Siberian Sea) is critical for understanding the region's pre-Cretaceous history.

Physiographic features of the oceanic realm typically used in paleogeographic reconstructions such as seafloor morphology (i.e., rift valleys) and magnetostratigraphy have limited applicability beyond the Eurasian and Canada Basins, because much of the oceanic and transitional crust that flanks the Amerasia Basin (Fig. 1A) has been obscured by magmatism and/or thick sedimentary cover (e.g., Døssing et al., 2013; Oakey and Saltus, 2016). This, in turn, has created a need to understand the geology of the continental fragments that make up the basin margins to establish and test paleogeographic and kinematic models of pre-opening configuration and ocean basin formation. Circum-Arctic correlations have leveraged relationships such as facies trends, the ages and locations of orogenic belts, and geochronology to link various crustal elements now separated by the modern Arctic Ocean basin and further understand their prior, obscured histories (e.g., Silberling et al., 1994; Colpron and Nelson, 2009; Moore and Box, 2016; Miller et al., 2018; McClelland et al., 2021).

St. Lawrence Island is a remote and poorly studied locale between Chukotka and Arctic Alaska in the Bering Strait region (Fig. 1B) that contains exposures of sedimentary rocks of Devonian–Triassic age, best exposed in the east-central part of the island (Figs. 2A and 2B; Patton and Csejtey, 1980), which were the focus of this study. Despite St. Lawrence Island's proximity to the Arctic Alaska–Chukotka microplate, the island's stratigraphy has never been conclusively linked with the main parts of the Arctic Alaska–Chukotka microplate: Chukotka and Wrangel Island in eastern Russia, and the Seward Peninsula, the Lisburne Peninsula, and the Brooks Range in Alaska (e.g., Lawver et al., 2002; Miller et al., 2010). Although many parts of the Arctic Alaska–Chukotka microplate are underlain by Neoproterozoic basement (e.g., Amato et al., 2009, 2014; Gottlieb et al., 2018), the oldest rocks known on St. Lawrence Island are Paleozoic sedimentary rocks; no basement exposures have been mapped (Patton and Csejtey, 1980).

This is the first study to use detrital zircon (DZ) ages from St. Lawrence Island to address correlations with the Arctic Alaska–Chukotka microplate and other parts of the Arctic region. Our investigation included lithologic and DZ data from western Brooks Range strata that appear to be correlative in age and depositional environment to stratigraphic successions studied on St. Lawrence Island. These new lithologic and DZ data from St. Lawrence Island and the western Brooks Range establish ties between St. Lawrence Island and several parts of the Arctic Alaska–Chukotka microplate, including the western Brooks Range and Lisburne Peninsula in northwestern Alaska and Wrangel Island in the East Siberian Sea (Fig. 1B), and the data set documents similar DZ age distributions to Carboniferous-age strata in the New Siberian Islands and Triassic strata in Arctic Canada (Fig. 1A). Despite St. Lawrence Island being a small and remote island disconnected from the larger landmasses on either side of the Bering Strait, these results, which conclusively link the island with the Arctic Alaska–Chukotka microplate and demonstrate multiple discrete sediment source regions for St. Lawrence Island Devonian–Triassic stratigraphic successions, are important contributions to the growing database of stratigraphic studies and DZ characteristics of circum-Arctic terranes.

The exposures of circum-Arctic bedrock geology have cryptic relationships to one another and to the cratonal landmasses of Laurentia, Baltica, and Siberia (Fig. 1), making paleogeographic reconstructions difficult. Previous efforts to understand the origins and tectonic transport of Arctic Alaska–Chukotka microplate segments have been based on faunal evidence (e.g., Blodgett et al., 2002; Strauss et al., 2013; Dumoulin et al., 2014a), paleomagnetic studies (e.g., Stone, 1989), stratigraphic and other geographic correlations (e.g., Embry, 2000; Toro et al., 2004; Grantz et al., 2011), and geophysical studies of oceanic and transitional lithosphere (e.g., Shephard et al., 2013; Houseknecht and Connors, 2016; Zhang et al., 2019; Døssing et al., 2020). DZ geochronology is another powerful tool that has been used to test lithostratigraphic and biostratigraphic correlations of Arctic Alaska–Chukotka microplate strata (e.g., Miller et al., 2006, 2010, 2011; Amato et al., 2009; Johnson et al., 2016; Strauss et al., 2017). However, St. Lawrence Island has been omitted from a well-known terrane map of the region (Silberling et al., 1994), possibly because its remote location in the Bering Sea makes along-trend correlations with better-studied circum-Arctic terranes difficult (Fig. 1B).

The various circum-Arctic reconstructions (e.g., Lawver and Scotese, 1990; Lawver et al., 2002; Amato et al., 2015; McClelland et al., 2021) allow for inferences about DZ source regions based on paleogeographic constraints at relevant time intervals. Observations and explanations of the presence or absence of certain ages in Arctic DZ samples have led to an empirical understanding of the ways in which the source regions with those ages inform the geologic history of the Arctic. A simple yet seminal example is how differences in early Mesoproterozoic DZ ages can be correlated to their derivation from Baltican versus Laurentian source regions based on the presence or absence of 1.61–1.49 Ga DZ populations during the North American magmatic gap (e.g., Colpron and Nelson, 2009; Miller et al., 2010, 2011). However, a complicating implication of the collision between Baltica and Laurentia during the Paleozoic Caledonian orogeny is that intra-orogen depositional settings that likely extended into the Arctic (e.g., fig. 4 inBlakely, 2021) could have mixed sediment provenance from both Baltican and Laurentian source regions, thus obfuscating the magmatic gap signal (or lack thereof). Additionally, recycling (i.e., intrabasinal reworking) of mid-Paleozoic strata with both Laurentian and exotic age distribution characteristics is well documented in Carboniferous and younger strata on the northern Laurentian margin of the Arctic (e.g., Gottlieb et al., 2014; Midwinter et al., 2016; Hadlari et al., 2018; Beauchamp et al., 2019), demonstrating the importance of integration of traditional petrographic and stratigraphic insights with DZ studies. Despite these and numerous other complexities regarding source region ambiguity, distinctive DZ age populations are nonetheless important signals in the sedimentary record, which, in conjunction with their depositional framework(s), have been leveraged to trace the influxes, distribution, and reworking of specific DZ populations in the Arctic to constrain provenance and paleogeography.

For Neoproterozoic and younger DZs, several distinctive age populations that coincide with the timing of major orogenic events in the Arctic have been indicated as key signals of past tectonics and paleogeography, although the ages are nonunique in terms of tectonic events. For example, Ordovician–Silurian zircons with ages ca. 510–435 Ma are nearly ubiquitous constituents of sedimentary deposits in the Arctic realm (including several samples from this study) and have been primarily associated with the Caledonian orogen (Caledonides, Fig. 1A; e.g., Gee et al., 2006; Lorenz et al., 2008; Miller et al., 2010; O'Brien and Miller, 2014; Brumley et al., 2015; O'Brien et al., 2016; Ershova et al., 2018a, 2018b). However, isotopic fingerprinting of Arctic DZs that have “Caledonian” (i.e., Ordovician and Silurian) ages has shown in some cases that such zircons were derived from relatively isotopically primitive magmatic settings (e.g., oceanic arcs) that existed beyond the range of continental-derived clastic influxes (McClelland et al., 2021). Thus, these zircons have more primitive geochemical signatures, which are inconsistent with having a Caledonian orogen source despite overlapping in age, exemplifying the potential confusion that could be caused by referring to them as “Caledonian” solely on the basis of their ages.

Likewise, late Neoproterozoic–Cambrian–age zircons (ca. 710–550 Ma) in Arctic strata are typically associated with the Timanian orogeny, which affected northern Baltica (e.g., Lorenz et al., 2008; Miller et al., 2011; Anfinson et al., 2012; Ershova et al., 2015a), as magmatism throughout that age range has been well documented in the Timanides (despite ages older than ca. 615 Ma predating the onset of orogeny; e.g., Larionov et al., 2004; Kuznetsov et al., 2007), and basement-cover age relationships have been well established (e.g., Pease and Scott, 2009; Kuznetsov et al., 2010; Miller et al., 2011). However, unrelated or poorly understood source terranes with igneous rocks of those ages also exist in the circum-Arctic (Fig. 1A), including: (1) the Seiland igneous province in northern Norway (alkaline intrusions at ca. 580–560 Ma; Roberts et al., 2006, 2010); (2) Chukotka in Russia and the Seward Peninsula of Alaska, which have granitic rocks at ca. 710–560 Ma (see Regional Geology section); and (3) the Central Taimyr belt, which has ca. 750–600 Ma ophiolitic rocks of uncertain paleogeographic origin (Pease and Vernikovsky, 2000; Vernikovsky et al., 2004; Priyatkina et al., 2017). Many siliciclastic rocks in the Arctic (including most from this study) contain broad multimodal distributions of Neoproterozoic–Cambrian DZ ages and also multimodal Ordovician–Silurian DZ age distributions, which are explained as resulting from mixing of DZs from multiple sources in an orogenic setting (Anfinson et al., 2012), or the postorogenic reworking of such strata (Gottlieb et al., 2014).

Permian–Triassic DZ ages are important in the Arctic because they may have been derived from Uralian orogen sources (Miller et al., 2006, 2013; Omma et al., 2011; Anfinson et al., 2016; Gilmullina et al., 2021a, 2021b) or from postulated Permian–Triassic arc complexes that fringed parts of the northern margin of Pangea (Hadlari et al., 2018; Alonso-Torres et al., 2018). By Triassic time, the Uralian orogen, which sutures Baltica and Siberia (Fig. 1A), existed as a drainage divide between these two cratons (e.g., Miller et al., 2013). Sediment transported along strike of and beyond the Urals into the Arctic region (i.e., northern Pangea margin; fig. 1 inMiller et al., 2013) could thus contain a mix of Baltica- and Siberia-derived zircons, complicating original provenance interpretations. The Uralian orogen was invoked by Gilmullina et al. (2021a) as the source for sediments transported across the entire Arctic region, overfilling an immense intracratonic basin known as the Greater Barents Sea Basin and into Chukotka and Alaska (Gilmullina et al., 2021b). Sources of Triassic zircons from felsic magmatism postulated to be associated with the Siberian Traps large igneous province could also have played a role in the provenance of sediment (Miller et al., 2006). Additionally, the generally widespread occurrence of near-depositional-age zircons in Permian–Triassic strata across the circum-Arctic regions, whether due to sourcing from the Urals or elsewhere, is a useful diagnostic tool in assessing depositional age insights about Arctic strata with metamorphic histories that have obscured lithostratigraphic and/or biostratigraphic approaches to determine stratigraphic age. Furthermore, the geographic extent of Permian–Triassic strata of known age, but lacking in near-depositional-age DZs (i.e., youngest DZs are considerably older than the depositional age), is an important observation in Arctic paleogeographic reconstructions, especially in relation to the Arctic Alaska–Chukotka region (e.g., Miller et al., 2006; Beranek et al., 2010; Gottlieb et al., 2014).

The Arctic Alaska–Chukotka microplate or microcontinent (Fig. 1A; e.g., Fujita, 1978; Hubbard et al., 1987; Silberling et al., 1994; Patrick and McClelland, 1995; Amato et al., 2009; Miller et al., 2010) consists of the Seward Peninsula of northwestern Alaska, the Brooks Range and North Slope of northern Alaska, the Chukotka Peninsula of eastern Russia, and islands and shelfal regions in the East Siberian and northern Bering Seas. St. Lawrence Island of Alaska (Fig. 1B) has been variably omitted (Silberling et al., 1994) and included (Amato et al., 2009) in the delineation of the Arctic Alaska–Chukotka microplate. Proterozoic basement of the Arctic Alaska–Chukotka microplate has igneous rocks ranging from ca. 1.0 Ga to the latest Neoproterozoic (see summary in Amato et al., 2014), including: (1) Wrangel Island, which has 711–620 Ma granitic and volcanic rocks (Gottlieb et al., 2018); (2) Seward Peninsula, which has 875–850 Ma meta-igneous rocks, 680–670 Ma intrusions with presumably older undated metasedimentary country rocks, 565 Ma orthogneiss, and 540 Ma gabbro (Amato et al., 2009, 2014); (3) Chukotka, which has orthogneisses ranging from 661 to 612 Ma (Amato et al., 2009, 2014; Gottlieb et al., 2018); and (4) the Brooks Range, which has a 968 Ma intrusion (Amato et al., 2014).

The Arctic Alaska–Chukotka microplate lower stratigraphic column contains two temporally discrete sedimentary megasequences, termed Franklinian and Ellesmerian (Lerand, 1973; for synopsis, see Gottlieb et al., 2014), separated in northern Alaska by an important regional unconformity surface referred to as the pre-Mississippian unconformity (e.g., Moore et al., 1994). The unconformity is underlain by penetratively deformed (“Franklinian”) Early Devonian and older strata and is beneath suprajacent Mississippian to Triassic (“Ellesmerian”) strata (e.g., Moore et al., 1994) across the North Slope and in parautochthonous parts of the Brooks Range. Detailed stratigraphic studies in the northeastern Brooks Range have documented that the older succession consists of Tonian–Lower Devonian strata, beginning with rift-related sedimentation in Tonian–Cryogenian time associated with Rodinia breakup, and containing an angular unconformity of Late(?) Ordovician–Early Devonian age that predates the pre-Mississippian unconformity (Strauss et al., 2019). Stratigraphic relationships also show that the penetrative deformation of (“Franklinian”) basement strata occurred during late Early Devonian time (Lane, 2007; Anderson and Meisling, 2021). Pre-Mississippian strata were exposed due to regional uplift of the northern Alaska part of the Arctic Alaska–Chukotka microplate in the Middle and Late Devonian (Moore et al., 1994). Thus, during Devonian–Mississippian time, multiple tectonic events are recognized in the Arctic Alaska–Chukotka microplate that created widespread unconformity surfaces (and associated erosion of pre-Mississippian strata).

The latter pre-Mississippian unconformity was most likely due to rift opening of the Angayucham Ocean to the south, which caused substantial amounts of exhumation and associated development of Ellesmerian highlands to the north (see Homza et al., 2020) and created a near-peneplain erosional surface at the pre-Mississippian unconformity in the Late Devonian (Moore et al., 1994). The sediment generated by this event was initially deposited in rift basins as Hunt Fork Shale, Noatak Sandstone, and Kanayut Conglomerate (Endicott Group) strata. As these basins filled, the entire margin was thermally subsiding, leading to initiation of the postrift Ellesmerian passive-margin megasequence in the earliest Mississippian. Initially, Ellesmerian megasequence sediment was eroded from the highlands source region and transported southward via fluvial to shallow-marine distributory/depositional systems, which onlapped progressively northward due to subsidence-driven marine transgression (Moore et al., 1994). Stratigraphic relations across the older unconformity show that the penetrative deformation (Romanzoff orogeny) occurred during early Devonian time (Lane, 2007). Across most of northern Alaska, the (younger) pre-Mississippian unconformity is overlain by a thin sequence of mainly fluvial conglomerate and sandstone, which is called the Kekiktuk Conglomerate (e.g., Brosge et al., 1962; LePain et al., 1994). These strata are widespread across northern Alaska and represent the onset of a latest Devonian to Jurassic south-facing passive-margin sequence. The clastic deposits of the Ellesmerian megasequence were likely derived from Franklinian rocks in the Ellesmerian highlands that existed along the pericratonic flank of the Arctic Alaska–Chukotka microplate and that are now found under the North Slope and northeastern Brooks Range due to the northward transgression of the Ellesmerian succession in the late Paleozoic and early Mesozoic (Moore et al., 1994).

The geology of St. Lawrence Island consists chiefly of voluminous Cretaceous granites and volcanic rocks and a large Quaternary shield volcano (Patton and Csejtey, 1980; Patton et al., 2011; Fig. 2A). Triassic and older sedimentary rocks are exposed on the northwestern and eastern parts of the island and were divided by Patton and Csejtey (1980) into two successions (Figs. 2A and 2B). The first is a Devonian–Triassic carbonate succession, which includes Devonian, Mississippian, and Triassic rocks (map units Dd, Ml, and Trs, respectively; Patton and Dutro, 1969; Ormiston and Fehlmann, 1969). The second is a Mississippian(?)–Triassic clastic succession, which consists mainly of Triassic mudstone and graywacke turbidites (unit TrPs of Patton and Csejtey, 1980), but which also has an arkosic lower member of uncertain age. Both successions experienced contact metamorphism but have not undergone regional metamorphism; original sedimentary textures are locally preserved. Patton et al. (2011) suggested that the two successions were parts of distinct “terranes” but also noted that contacts between the two successions locally appear to be depositional. If the two successions represent different facies, then their relative stratigraphic positions are unknown, given the poor exposures and abundant faulting (Patton and Csejtey, 1980). Lithologies and fossil age control of the Triassic and older sedimentary successions on St. Lawrence Island were summarized in Patton and Csejtey (1971, 1980), Till and Dumoulin (1994), and Patton et al. (2011).

Paleozoic and Mesozoic sedimentary rocks of the western Brooks Range in the Arctic Alaska–Chukotka microplate are exposed in a series of discrete allochthons (Fig. 3; Mayfield et al., 1988; Young, 2004) that were emplaced during Jurassic–Early Cretaceous north-directed shortening associated with arc-continent collision (e.g., Moore et al., 1994). The Brookian orogeny occurred during collision of an oceanic arc against Arctic Alaska's south-facing margin (present-day reference frame), closing the Angayucham Ocean basin; the thrust-juxtaposed remnants of this event are referred to as the Angayucham terrane (e.g., Wirth et al., 1993; Silberling et al., 1994). The allochthons of the Brookian orogen, including the Angayucham terrane and the parautochthonous passive margin of Arctic Alaska, are chiefly distinguished by differences in their Devonian and Mississippian stratigraphy (Mayfield et al., 1988; Young, 2004; Dumoulin et al., 2004, 2006, 2014b). The structurally lower allochthons originated closer to the Brookian parautochthon, as indicated by sedimentary facies becoming progressively more distal moving structurally higher in the allochthon stack (e.g., Mayfield et al., 1988; Moore et al., 2015, and references therein; Fig. 3). The parautochthon underlies all of northern Alaska at least as far south as the Doonerak Fenster in the central Brooks Range (Fig. 1B). In the western Brooks Range, the structurally lowest thrust sheet is the Endicott Mountains allochthon, which is characterized by thick siliciclastic strata of Devonian–Mississippian age (Mayfield et al., 1988; Moore et al., 1994; Dumoulin et al., 2004). The next highest sheet is the Picnic Creek allochthon, which includes Carboniferous rocks representative of deep-water facies (carbonate turbidites, radiolarian chert) with minimal siliciclastic detritus (e.g., Dumoulin et al., 2004). Above this, there is the Kelly River allochthon, which is the only allochthon that has a carbonate-dominated, rather than siliciclastic-dominated, Devonian section (e.g., Dumoulin et al., 2004). The Kelly River allochthon contains a shallow-water Carboniferous carbonate section (Lisburne Group) that includes a volumetrically significant component of silt-sized and sand-sized detrital quartz (Mayfield et al., 1988; Dumoulin and Harris, 1992; Young, 2004; Dumoulin et al., 2004, 2006, 2014b). The Kelly River allochthon is overlain by the Ipnavik River allochthon, which includes Carboniferous strata similar to those of the Picnic Creek allochthon in terms of composition and inferred depositional setting (Dumoulin et al., 2004). The Nuka Ridge allochthon is the structurally highest (i.e., most distal) of the five Brooks Range allochthons that are principally composed of sedimentary rocks (Fig. 3B; Young, 2004). Its distinguishing feature is the Carboniferous Nuka Formation (Tailleur and Sable, 1963; Tailleur et al., 1973; Solie and Mull, 1991; Moore et al., 1997), a distinctive feldspathic limestone and sandstone unit that was one of the units sampled as part of this study.

Our study of St. Lawrence Island strata used archival rock samples collected by W. Patton and B. Csejtey of the U.S. Geological Survey during field work there in 1968–1971, as well as thin sections made from these samples, archived field notes, geologic maps, and photographs. These rocks are reposited at the Geologic Materials Center, Alaska Division of Geologic and Geophysical Surveys, in Anchorage, Alaska; all other archival materials are stored in the Technical Data Unit of the Alaska Science Center, U.S. Geological Survey, Anchorage, Alaska. Petrographic descriptions were based on examination of 113 thin sections (see list in Dumoulin et al., 2022a). Chronostratigraphic divisions followed Walker et al. (2018).

Ten archival samples from St. Lawrence Island, three from the Devonian–Triassic carbonate succession and seven from the Mississippian(?)–Triassic clastic succession, were analyzed for DZ U-Pb ages at the Arizona LaserChron Center (ALC) using laser-ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) techniques (Table 1). These small hand samples (typically ~300 g) were processed for heavy mineral separation at New Mexico State University. First, they were disaggregated in a jaw crusher and disc grinder to fine sand size. The fraction finer than 60 mesh sieve (<250 μm) was washed in a 1 L beaker, and the clay-sized fraction was poured out. Dried sample was put in a Frantz magnetic separator to 1.5 A, and the nonmagnetic material was soaked in 10% acetic acid to remove carbonates and 3% H₂O₂ to remove remaining clays. Dried sample was then put in sodium polytungstate (2.85 g/cm³) and then methylene iodide (3.3 g/cm³). If pyrite was present, the sample was soaked in 15% nitric acid for 30 min on a hot plate to dissolve sulfides.

For ALC LA-ICP-MS analyses, samples were mounted in epoxy, polished, and imaged using standard methodology and analyzed with an Analyte G2 193 nm excimer laser (20 µm beam diameter) coupled with a Thermo Element 2 single-collector mass spectrometer at the ALC following the methods described in Pullen et al. (2018). Sample 91AMu-107 from the Nuka Ridge allochthon in the western Brooks Range was processed for mineral separation at Stanford University and analyzed in 2011 at ALC using a New Wave UP193HE Excimer laser coupled with a Nu Instruments high-resolution (HR) ICP-MS in multicollector mode. Details of the sample preparation and analysis are as described in Gottlieb et al. (2014).

For Brooks Range samples 04CP-48 from the Kelly River allochthon and 06TM-28 from the Nuka Ridge allochthon, mineral separation work and U-Pb LA-ICP-MS analyses were done by Apatite to Zircon, Inc., in 2012, with instrument work occurring at the Geoanalytical Laboratory at Washington State University. LA-ICP-MS analyses of sample DT-84–14/15 from the Nuka Ridge allochthon were also done at the Geoanalytical Laboratory at Washington State University by GeoSep Services in 2021 from remnants of two samples studied by Moore et al. (1997). Sample preparation procedures were similar to those reported in Bradley et al. (2009), Hults et al. (2013), and Moore et al. (2015). Epoxy wafers (~1 cm × 1 cm) containing zircon grains for LA-ICP-MS were polished manually using Al₂O₃ slurries to expose internal zircon grain surfaces. The polished zircon grain surfaces were washed in 5.5 M HNO₃ for 20 s at 21 °C to clean the grain surfaces prior to introduction into the laser system sample cell.

Samples 04CP-48 and 06TM-28 were analyzed using a Finnigan Element II magnetic sector ICP-MS equipped with a New Wave Nd-YAG 213 nm laser-ablation system, whereas sample DT-84–14/15 was analyzed using the New Wave UP-213 laser-ablation system in conjunction with an Agilent 7700x quadrupole ICP-MS. A 20 µm beam diameter was used for each sample.

Default ALC quality control and age picking cutoffs were used for data reduction. Rejection criteria used were: ²⁰⁶Pb/²⁰⁴Pb <200, ²⁰⁶Pb/²³⁸U percent error (1σ) >10%, ²⁰⁶Pb/²⁰⁷Pb percent error (1σ) >10%; and percent discordance >20% or <–5%. The ²⁰⁶Pb/²⁰⁷Pb and discordance filter criteria were not used for ages younger than 700 Ma. Analyses that met the rejection criteria were not reported in individual sample data tables and therefore were not included in the study. Systematic errors (2σ) of 0.8% for ²⁰⁶Pb/²³⁸U ratios and 0.5% for ²⁰⁶Pb/²⁰⁷Pb ratios were propagated into uncertainties for these ratios in all Element 2 analyses, and greater (2σ) uncertainties of 1.1% for ²⁰⁶Pb/²³⁸U ratios and 0.8% for ²⁰⁶Pb/²⁰⁷Pb ratios were propagated for Nu HR analyses. For ALC data, individual “best ages” are reported as ²⁰⁶Pb/²³⁸U ages for all ²⁰⁶Pb/²³⁸U age estimates younger than 900 Ma and as ²⁰⁷Pb/²⁰⁶Pb ages for all ²⁰⁶Pb/²³⁸U age estimates older than 900 Ma. For data collected by Apatite to Zircon, Inc., and GeoSep Services, individual “best ages” are reported as ²⁰⁶Pb/²³⁸U ages for all ²⁰⁷Pb/²⁰⁶Pb age estimates younger than 900 Ma and as ²⁰⁷Pb/²⁰⁶Pb ages for all ²⁰⁷Pb/²⁰⁶Pb age estimates older than 900 Ma. With the exception of sample 06TM-28, all analyses are reported in Part 4 of the Supplemental Material¹. Sample 06TM-28 had several analyses that yielded incalculable ratios for determining age results, and these were summarily excluded from the data table. Geochronology results are included as Supplemental Material and are also in Dumoulin et al. (2022a).

Several standard geochronology calculations and data summarizations were made from the individual samples. Maximum depositional ages (MDA) were calculated using the method of weighted mean age of the youngest group of three or more zircons with ±2σ age overlap, known as YC2s(3+), and calculated using DetritalPy (Sharman et al., 2018; see also Dickinson and Gehrels, 2009; Coutts et al., 2019). The maximum likelihood age (MLA; Vermeesch, 2021) was also calculated using Isoplot R (Vermeesch, 2018) to determine the youngest statistically valid age in each sample. A mean age result was calculated for the predominant ca. 2.05 Ga age mode observed in each sample from the St. Lawrence Island feldspathic subunit of the Mississippian(?)–Triassic clastic succession and the western Brooks Range Nuka Formation using IsoplotR (Vermeesch, 2018). All weighted mean and single zircon ages are reported at the ±2σ level. The mean square of weighted deviates (MSWD; Wendt and Carl, 1990) is reported for each mean age.

Kernel density estimation (KDE) plots were generated in Isoplot R (Vermeesch, 2018) using a kernel bandwidth of 20 (m.y.) and keeping bandwidth fixed for all samples shown in a given figure. Cumulative age distribution function plots were also generated using Isoplot R (Vermeesch, 2018).

The Devonian–Triassic carbonate succession makes up the Ongoveyuk terrane of Patton et al. (2011); see Figures 4A–4F and 5A–5F herein. It is best exposed along streams that drain a wave-cut platform on the eastern part of St. Lawrence Island, but it is also recognized in a few smaller exposures to the west (Fig. 2). Samples examined and analyzed for our study came from several outcrops of units Dd, Ml, and Trs in the eastern part of the island (Fig. 2; list of samples examined in Dumoulin et al., 2022a).

The succession begins with unit Dd (Devonian dolostone and dolomitic limestone; Patton et al., 2011), which may be as much as 1300 m thick (Fig. 4A). Thin section textures, best preserved in dolomitic limestones but visible in some dolostones, ranged from wackestone to grainstone; clasts were mainly skeletal grains and micritic peloids (Figs. 5A and 5B). Bioclasts observed in thin section included crinoid ossicles, ostracodes, gastropods, algae, foraminifers, and fragments of brachiopods and bryozoa (Figs. 4F and 5B). Fossil fragments in several thin sections were partly to completely micritized. Some samples are well layered, with millimeter-scale microbial lamination (Fig. 4E).

Conodonts indicate that unit Dd is at least as old as earliest Devonian and possibly as old as late Silurian (Ludlow–Lochkovian; A.G. Harris, U.S. Geological Survey, 1993, written commun.; Till and Dumoulin, 1994). Chonetid brachiopods from the lower part of the succession are similar to (and possibly conspecific with) Emsian (late Early Devonian) specimens from the northeastern Brooks Range (Blodgett, 2017). Tabulate corals (including Favosites sp. and thamnoporoid forms) higher in the section are probably Middle or early Late Devonian (Givetian or Frasnian) forms (Patton and Dutro, 1969). Stromatolites occur in the lower part of Dd; other fossils identified to species or genera from Dd include several types of gastropod, pentamerid brachiopods, ostracodes, and diverse stromatoporoids (Ormiston and Fehlmann, 1969). The fauna as well as features such as peloids, micritized bioclasts, and microbial layering denote shallow to very shallow, locally hypersaline, depositional settings.

The Mississippian section (unit Ml, Mississippian limestone; Patton et al., 2011) is at least 400 m thick and abundantly fossiliferous (Fig. 4A). The lower part is silty, locally dolomitic limestone (Fig. 5C) interbedded with very fine–to medium-grained sandstone made up of various proportions of subrounded to angular monocrystalline quartz and chert grains ± carbonate clasts and bioclasts, with carbonate or silica cement (Fig. 4D). Some chert clasts are cut by thin quartz veinlets, and others contain siliceous sponge spicules. Subordinate grain types in these sandy beds include shale clasts, mica, tourmaline, and zircon. Siliciclastic detritus decreases upward, and much of the section consists of skeletal grainstone (Figs. 4C and 5D) and packstone, with locally abundant chert nodules. The section becomes muddier toward the top (Fig. 4A). Bioclasts are mainly echinoderms (crinoids) and diverse bryozoans (Fig. 4C), with lesser algal and brachiopod fragments, foraminifers (Fig. 5D), ostracodes, and sponge spicules. A few samples contained peloids, micritic intraclasts, and/or partly micritized skeletal grains. Definitively dated fossils include Visean foraminifers and Meramecian lithostrotionid corals (Patton and Dutro, 1969). Conodonts from the upper part of the section are late Meramecian–early Chesterian (Lane and Ressmeyer, 1985). Faunal and sedimentologic evidence indicates that the Ml unit accumulated in an array of shallow-water, mostly normal-marine settings with an overall deepening-upward trend.

The uppermost part of the carbonate-dominated succession is ~120 m thick (Fig. 4A) and consists of thinly interbedded chert, siltstone, shale, and fine-grained limestone, with limestone predominant near the top of the unit (Trs, Triassic shale, limestone, and chert; Patton et al., 2011). Cherty beds contain locally abundant radiolarians and lesser siliceous sponge spicules. Limy beds are lime mudstones and pelecypod wackestones-packstones with common calcitized radiolarians (Fig. 5E) and rare echinoderm fragments; some pelecypod shell fragments (Monotis sp.) contain microborings (Fig. 5F). The lower half of Trs has not been dated, but the upper part contains a succession of pelecypod bivalves indicating that upper Middle and Upper Triassic strata are present; bivalves of middle Ladinian (Daonella frami), Carnian (Halobia cf. H. superba–H. zitteli), and middle and late Norian (Monotis cf. scutiformis pinensis; M. cf. M. subcircularis; Fig. 4B) age have been identified (N. Silberling, U.S. Geological Survey, 1968, written commun.; Patton and Dutro, 1969; Ormiston and Fehlmann, 1969; Blodgett, 2017).

Map unit TrPs (Triassic–Permian graywacke, grit, and shale) of Patton and Csejtey (1980) and Patton et al. (2011) was assigned by those workers to a “graywacke succession” and was estimated by them to be as much as 800 m thick (Fig. 6). It is spatially associated with unit TrPg (Triassic–Permian gabbro and diabase), which consists mainly of coarse-grained mafic igneous rocks and subordinate tuff, basalt, graywacke, and mudstone. Two K-Ar age determinations on gabbro in TrPg produced Triassic ages (221 ± 7 Ma and 244 ± 7 Ma; Patton and Csejtey, 1980; Patton et al., 2011). Together, these two units constitute the Tomname terrane of Patton et al. (2011). Like the carbonate succession, the graywacke succession is best exposed on the eastern part of St. Lawrence Island, but it is also found in several small rubble exposures to the west; we examined and analyzed samples from both eastern and western exposures (Fig. 2; Dumoulin et al., 2022a). Published reports (Patton and Csejtey, 1980; Patton et al., 2011) and field notes (W. Patton, 1970, 1971, written commun.) indicate that the graywacke succession encompasses two subunits that differ strikingly in texture and composition. Thin section analyses and new DZ data confirm this distinction, and the two parts of the succession are discussed separately here.

Strata assigned to the lower part of the graywacke succession by Patton and Csejtey (1980) and Patton et al. (2011) are not graywacke rocks per se; instead, they consist of highly distinctive arkosic to subarkosic, medium-to coarse-grained sandstone and pebble conglomerate interbedded with impure clastic limestone (Figs. 6F–6K). To avoid confusion, these strata are herein called the “feldspathic subunit.” Beds are massive, up to 1 m thick, and locally graded. Sorting is generally moderate, and many clasts (especially in coarser sandstone and conglomerate) are rounded to subrounded (Figs. 6F, 6G, and 6I). Sandstone samples contain 15%–50% plagioclase ± potassium feldspar and abundant monocrystalline and polycrystalline quartz clasts in a matrix of calcite or dolomite cement. Feldspar grains have a variety of features, including tartan (polysynthetic) twinning (Fig. 6J) and perthitic, antiperthitic, and myrmekitic textures. Some feldspar grains are largely altered to sericite and/or calcite. Many quartz grains appear highly strained, and some are embayed. Plutonic (granitic) clasts are locally common, as are sedimentary clasts (fine-grained limestone, mudstone, and phosphate; Figs. 6G and 6I) and bioclasts (mainly echinoderm debris; Fig. 6H). Clasts and lenses of fine-grained carbonate containing calcitized radiolarians and calcareous sponge spicules occurred in several samples. Other subordinate components include volcanic lithic clasts, chert, and zircon. Rugose coral fragments of probable Mississippian–Permian age (W.J. Sando, 1971, written commun.), found in a feldspathic carbonate conglomerate interval, suggest a late Paleozoic age for the feldspathic subunit (Fig. 6).

The main part of the graywacke succession, herein called the “graywacke subunit,” consists of at least several hundred meters of carbonaceous mudstone with subordinate (15%–25%) interbeds of siltstone to medium-grained carbonaceous sandstone, rare shale pebble conglomerate, and local intervals of thin-bedded, varicolored (black, gray, and green) radiolarian chert (Patton and Dutro, 1969; Patton and Csejtey, 1980; Figs. 6A–6E). Many silty to sandy beds are graded, and sole markings are common. Sandstone in the graywacke subunit differs in both texture and composition from the arkosic, carbonate-bearing beds in the feldspathic subunit. Graywacke subunit sandstone is mostly fine to very fine grained and heterolithic, lacks carbonate, and contains abundant phyllosilicate matrix; clasts are mainly angular to subangular (Figs. 6B–6D). Clast types include monocrystalline quartz, plagioclase feldspar, notable mica (biotite, white mica, and chlorite), and a variety of volcanic, sedimentary, and metamorphic lithic grains (Fig. 6D); in contrast, micas and metamorphic lithic grains are absent from the feldspathic subunit samples. Sedimentary structures indicate a turbidite origin for the graywacke subunit (Patton and Dutro, 1969). Chert interbeds yielded radiolarians of probable Triassic age (Patton et al., 2011).

In their initial description of the St. Lawrence Island Devonian–Triassic carbonate succession, Patton and Dutro (1969) suggested a correlation with coeval strata in northern Alaska. Studies of Brooks Range stratigraphy, sedimentology, and structure, including new and previously published data synthesized below, allow us to propose a more precise correlation. The Kelly River allochthon (Figs. 3 and 7) provides the best match for the stratigraphy of the Devonian–Triassic carbonate succession on St. Lawrence Island. Five plates at three structural levels are recognized within the Kelly River allochthon (Fig. 7A; Young, 2004); the lithofacies, biofacies, and overall stratigraphy of the Paleozoic through Triassic section differ somewhat from plate to plate (e.g., Dumoulin et al., 2006). Devonian and Mississippian strata in the Kelly River allochthon have been most studied and best dated in exposures of the Eli and Kelly plates in the southwestern Brooks Range (Baird Mountains and DeLong Mountains quadrangles; Dumoulin and Harris, 1992; Dumoulin et al., 2004, 2006). Permian–Triassic strata in the Kelly River allochthon are less well known but have been investigated at several localities in the Wulik Peaks and Kelly plates in the western Brooks Range (DeLong Mountains and Misheguk Mountain quadrangles; Dumoulin et al., 2006, 2011).

The Devonian section in the Kelly River allochthon is thick, contains abundant dolostone, spans the Devonian Period, and was deposited largely in shallow-water to very shallow-water settings (Figs. 7A and 7H–7J). In the Eli plate in the northwestern Baird Mountains (Fig. 7H), the basal unit is mainly dolostone, >640 m thick, and has abundant micritic peloids and fenestral fabric (Fig. 7I), as well as microbial layering (Fig. 7J), which are characteristic of supratidal to shallow subtidal environments (Dumoulin and Harris, 1992). Conodonts indicate an age of late Early to early Middle Devonian (Eifelian–Givetian) for the upper part of the dolostone unit; equivalent strata to the west and possibly correlative rocks to the southeast (Baird Group) are as old as Early Devonian (Emsian) and middle and late Silurian (Wenlock and Ludlow), respectively (Dumoulin et al., 2006; Till et al., 2008). Younger Devonian rocks in the Kelly River allochthon (Eli limestone, 165 m thick) are limestone and dolostone, locally argillaceous, that contain Late Devonian (Frasnian and Famennian) conodonts (Dumoulin and Harris, 1992). Other fossils in the Devonian section include stromatoporoids, corals, brachiopods, ostracodes, gastropods, echinoderms, bryozoans, foraminifers, and calcispheres (Dumoulin and Harris, 1992).

Mississippian strata in the Kelly River allochthon are abundantly fossiliferous and include a siliciclastic-rich lower unit, the Utukok Formation (one of the subjects of this study), overlain by limestone and dolostone of the Kogruk Formation. In the Eli plate, the Mississippian section is as much as 500 m thick (Fig. 7D; Dumoulin and Harris, 1992; Dumoulin et al., 2004). The Utukok Formation here consists of skeletal wackestone to grainstone interbedded with very fine–to fine-grained, monocrystalline quartz ± carbonate sandstone (Figs. 7F and 7G). Sandy beds make up ~40% of the section and are plane-to cross-laminated with silica or calcite cement. Grains are subangular to subrounded. Carbonate clasts include fine-grained limestone, bioclasts (Fig. 7F), and dolomite rhombs; other clasts are chert, feldspar, mudstone, and tourmaline. The Kogruk Formation is mainly skeletal wackestone to grainstone (Fig. 7E) with locally abundant chert nodules that deepens upward into thinly interbedded spiculitic limestone and shale. Fossils in both formations are diverse and include echinoderms, bryozoans, foraminifers, brachiopods, ostracodes, corals, and algae (Fig. 7E); bioclasts in some samples are partly to completely micritized. Conodonts indicate that the Utukok Formation in the Eli plate is Early Mississippian (Kinderhookian to middle Osagean [Tournaisian]) and the Kogruk is Middle Mississippian (late Osagean to early Chesterian [Visean]; Fig. 7).

The Kelly River allochthon succession is capped by a thin, chert-rich interval, the Etivluk Group (~40–80 m thick; Curtis et al., 1984). The lower part of the Etivluk Group, the Siksikpuk Formation, is mainly chert and shale that contain Permian radiolarians in the Kelly River allochthon (Fig. 7; Dumoulin et al., 2011); equivalent strata in other allochthons yielded Pennsylvanian and Permian fossil ages (Dumoulin et al., 2004, 2006; Young, 2004). The upper part of the Etivluk Group, the Otuk Formation, is radiolarian chert and shale, interbedded with limestone near the top of the unit. Monotid bivalve wackestone-packstone (Figs. 7B and 7C), identical in age and microfacies to that in the St. Lawrence Island carbonate succession, contain locally abundant calcitized radiolarians and bored bivalve fragments. Ladinian and Carnian radiolarians and Carnian(?) and Norian pelecypods constrain the age of the Otuk Formation in the Kelly River allochthon (Figs. 7A and 7B; Curtis et al., 1984; Dumoulin et al., 2011). In other allochthons (e.g., the Endicott Mountains allochthon), the Otuk Formation contains a succession of Ladinian, Carnian, and Norian pelecypods (e.g., Kelly et al., 2007) identical to that documented in the St. Lawrence Devonian–Triassic carbonate succession.

The Nuka Ridge allochthon is the structurally highest of the five Brooks Range allochthons that are made up mostly of sedimentary rocks (Fig. 3B; Young, 2004). Its distinguishing feature is the Carboniferous Nuka Formation (Tailleur and Sable, 1963; Tailleur et al., 1973; Solie and Mull, 1991; Moore et al., 1997), a distinctive feldspathic limestone and sandstone unit that has striking similarities to the feldspathic subunit of the graywacke succession on St. Lawrence Island.

The Nuka Formation (Fig. 8) consists of variously glauconitic and arkosic limestone and arkosic to subarkosic sandstone exposed in widely scattered small outcrops through the western and central Brooks Range (Curtis et al., 1984; Mayfield et al., 1988; Young, 2004). The maximum thickness of the Nuka Formation is ~250 m, and it formed in a range of shallow-to deep-marine (and nonmarine?) settings (Solie and Mull, 1991; Moore et al., 1997; Young, 2004). Strata from the type section at Nuka Ridge in the west-central Brooks Range (loc. 1, Fig. 1B; Fig. 8F) are locally glauconitic and consist of fossiliferous limestone, massive to medium-bedded, fine-to very coarse–grained arkosic sandstone, and subordinate granule conglomerate (Moore et al., 1997). Megafossils at this locality include echinoderms, brachiopods, and bryozoans. Shallow-marine sedimentary features, such as herringbone cross-stratification and low-angle, inclined planar lamination, are common (Moore et al., 1997). Nuka strata from Kikiktat Mountain, in the central Brooks Range (Figs. 8A–8E), were described in detail by Solie and Mull (1991) and consist of arkosic limestone and sandstone with interbedded shale. Arkosic beds are graded, contain Bouma Ta–Td sequences, and likely formed as turbidites (Figs. 8B and 8C). Sandy layers contain potassium feldspar, plagioclase, quartz, and calcareous bioclasts including crinoid fragments (Figs. 8D and 8E). A section of the Nuka Formation in the northwestern Noatak quadrangle includes bioclastic grainstone that contains common grains of quartz, plagioclase, microcline, and glauconite, as well as minor phosphate (Dumoulin et al., 2006). Studies of conodonts and foraminifers indicate that the depositional age of the Nuka Formation is Middle Mississippian–Early Pennsylvanian (late Meramecian–early Atokan; Curtis et al., 1984; Young, 2004; Dumoulin et al., 2006, and references therein).

Three samples of Mississippian quartz arenite (Ml unit of Patton et al., 2011) were analyzed for DZ U-Pb ages (Table 1; Fig. 9). Sample 71APa-231 (DZ ages analyzed: n = 110) was collected in the southeast part of the island (Fig. 2A) and consists of very fine-grained sandstone with subangular to subrounded grains of quartz, lesser chert and feldspar, trace tourmaline and white mica, and carbonate cement. Samples 70APa-100 (n = 107) and 70APa-120A (n = 109) were collected south of Tomname Lagoon (Fig. 2B); both are similar in texture and composition to 71APa-231, but they have silica, not carbonate, cement. Mississippian spiriferoid brachiopods were identified in the bed from which sample 70APa-120A was collected (J.T. Dutro, 1971, written commun.).

All three samples yielded similar DZ age distributions (Fig. 9) and MDAs of ca. 438–435 Ma (Llandovery, early Silurian; Table 1), which is ~100 m.y. older than the depositional age of the samples based on the faunal data. Although the samples showed slight variations in their age distributions, the three age distribution results have broad similarities that provide confidence that the data from these three data sets can be combined for interpretive purposes. These similarities include prominent modes in the youngest cluster of ages: Samples 70APa-100 and 70APa-120a have a mode at ca. 439 Ma; sample 71APa-231 has a mode at ca. 435 Ma, but its most prominent mode peak is at ca. 555 Ma. All three samples have minor ages between 950 and 600 Ma. All have modes at ca. 950 Ma and minor modes between 1650 and 950 Ma, and a paucity of ages older than 1800 Ma.

Four arkosic sandstone samples, from both eastern and western exposures of the feldspathic subunit of the Mississippian(?)–Triassic clastic succession (formerly the graywacke succession of Patton and Csejtey, 1980), were analyzed for DZ ages (Table 1; Fig. 10). The easternmost sample (71APa190–1; Fig. 2B) is medium to coarse grained with subequal amounts of subrounded to angular quartz and plagioclase feldspar, minor ferric dolomite rhombs, a few likely crinoid fragments, and several clasts of phosphate and shale. Its main group of ages (exclusive of outliers) has a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of 2057 ± 3 Ma (MSWD = 0.96; n = 97; Fig. 10A). Sample 71APa-195 (Fig. 2B) is similar in texture and composition but also includes a few volcanic and plutonic lithic clasts. Its main population has a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of 2057 ± 2 Ma (MSWD = 1.7; n = 199; Fig. 10B).

Two samples were collected from the northwestern part of the island (Table 1; Fig. 2A). Sample 70APa-31y is a medium-grained, poorly sorted, reddish-beige arkosic sandstone with rounded to subangular grains, made up of nearly equal amounts of quartz and feldspar with minor ferric dolomite (as clasts and cement) and a few chert clasts. Staining indicates plagioclase is slightly more abundant than potassium feldspar; some grains have tartan twinning. Its weighted mean ²⁰⁷Pb/²⁰⁶Pb age is 2055 ± 5 Ma (MSWD = 0.79; n = 28; Fig. 10C). Sample 70APa-31z is a grayish sandstone similar in composition and texture to sample 70APa-31y, but with calcite cement and bioclasts, including crinoid ossicles and several large shale chips. It has a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of 2056 ± 12 Ma (MSWD = 0.70; n = 5; Fig. 10D).

All four samples produced similar distinctively nearly unimodal DZ age distributions at ca. 2.06 Ga. Although these values generally also represent the MDA (Table 1), these ages are at least 1.7 b.y. older than paleontological constraints, which indicate that the unit must be late Paleozoic (Mississippian–Permian; Table 1; Fig. 10). Sample 71APa-190–1 yielded a single Mississippian (Visean) zircon (341 Ma) that we interpret to be relatively close to the stratigraphic age.

Three samples from the graywacke subunit of the Mississippian(?)–Triassic clastic succession (formerly the graywacke succession of Patton and Csejtey, 1980) were analyzed for DZ U-Pb ages (Table 1; Figs. 2 and 11). One sample (71APa-193; DZ ages: n = 117) was collected in the northeast part of the island (Fig. 2B). It is a very fine-grained sandstone made up of quartz, plagioclase, chert, metamorphic lithic clasts, biotite, chlorite, and white mica. A second sample (71APa-194; n = 97) came from an outcrop nearby; it resembles sample 71APa-193 in composition but is fine to medium grained and contains outsized silty mudstone clasts as well as some volcanic lithic grains. The third sample (70APa-103; n = 104), from an outcrop south of Tomname Lagoon (Fig. 2B), is similar in texture and composition to sample 71APa-193.

These three samples produced broadly similar DZ age distributions. The three samples all yielded Triassic MDAs (Table 1), ranging from ca. 217 Ma (Norian, Late Triassic) to ca. 252 Ma (earliest Triassic), consistent with the Triassic depositional age suggested for the graywacke subunit of the Mississippian(?)–Triassic clastic succession by the radiolarian data. However, the youngest zircon ages in all three samples overlapped in the ca. 217–215 Ma range. Because sample 71APa-193 lacked sufficient ages in that range, the MDA is necessarily older, but we suggest that the lithologic similarities and the similarities in the youngest zircon ages indicate that the three were likely deposited at about the same time and shared the same source. Note also that sample 71APa-193 was collected only 0.6 km from the location of sample 71APa-194, and both are in the same map unit. All samples contained abundant (30%–50%) Carboniferous–Triassic–age grains as well as an array of older Paleozoic and Proterozoic grains (Fig. 11). There were no differences in the age distributions among the three samples; the majority of ages were younger than ca. 700 Ma, and Proterozoic–Archean ages made up 30%–45% of each sample, although prominent Proterozoic or Archean modes were not recognizable in any of the age distributions. A combined KDE from all three samples had the most prominent mode at 299 Ma, and all of the major modes were between ca. 432 Ma and ca. 216 Ma (Fig. 11).

A sample of Utukok Formation sandstone from the Kelly plate of the Kelly River allochthon in the western Brooks Range was analyzed for DZs (sample 04CP-48; Table 1; Fig. 9; n = 100). The outcrop has medium (10–20 cm thick) beds with ripple lamination, and the sample is a very fine-grained quartz-carbonate sandstone with subordinate altered feldspar, chert, and white mica. Calcite is the cement, and there is patchy alteration of feldspar grains. Clasts include an outsized crinoid fragment.

This sample yielded a DZ age distribution with a latest Middle Devonian MDA of 402 Ma (Table 1), a prominent mode at 1584 Ma, a Silurian age mode (ca. 430 Ma), other modes at ca. 949 Ma and ca. 521 Ma, and an array of older Paleozoic and Proterozoic ages (Fig. 9). Most of the Proterozoic ages were younger than ca. 1.95 Ga, with only a single middle Paleoproterozoic zircon (ca. 2.04 Ga) that was close to the modal age of the feldspathic subunit samples.

Three samples from the Nuka Formation type locality (Nuka Ridge; Figs. 1 and 10) were studied. Sample 91AMu-107 is an arkosic limestone, very coarse grained to granular, with subangular to rounded quartz and potassium feldspar in a gray limestone matrix, and it contained large fragments of silicified brachiopod shells. The zircons from sample 91AMu-107 yielded a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of 2059 ± 2 Ma (MSWD = 0.46, n = 95; Fig. 10E). Only two grain ages were outside of this group, one at 1879 ± 67 Ma and another at 324 ± 8 Ma (Visean/Middle Mississippian); this youngest zircon age matches faunal data, suggesting that the Nuka Formation is Mississippian–Pennsylvanian in age (e.g., Dumoulin et al., 2006).

The two other samples were collected from ~6 km away. Sample DT-84–14 is an arkosic limestone with abundant calcite (42%), as clasts and cement, with ~25% each quartz and potassium feldspar. Sample DT-84–15 is an arkosic granule conglomerate made up mainly of potassium feldspar and quartz. Both samples contained minor amounts of plagioclase and lithic clasts, chiefly granite with rare quartzite and quartz–mica schist (for additional location, stratigraphic, and petrographic information, including point count data, see Moore et al., 1997). These samples (and a few others from the same locality) were analyzed by Moore et al. (1997), who reported multigrain thermal ionization mass spectrometry ages in the range of ca. 2078–2013 Ma. Samples DT-84–14 and DT-84–15 were combined (to increase yield; n = 220) and reanalyzed by LA-ICP-MS, separating the analyses into small and large grains (Table 1; Supplemental Material, ^{footnote 1}). The ages for each analytical group were essentially the same, and the data were imprecise (Fig. 10F) and had indications of Pb loss (Fig. 10G), so we interpret the discordia age (upper intercept) of 2058 ± 16 Ma (MSWD = 0.34; n = 220) to be the best age for this sample.

Another sample from the Nuka Formation was collected from the Poktovik Mountains in the western Brooks Range (Fig. 1). Sample 06TM-28 is a very poorly sorted, coarse-grained, arkosic sandstone with abundant quartz and potassium feldspar, lesser plagioclase, and traces of patchy calcite, mica, and glauconite. Many feldspar grains showed tartan twinning, and a few perthitic and myrmekitic grains were present. This sample gave an essentially unimodal zircon population with a ²⁰⁷Pb/²⁰⁶Pb weighted mean age of 2036 ± 6 Ma (MSWD = 0.38; n = 83; Table 1; Fig. 10). This age matches the weighted mean age of the DT-84–14/15 samples, but it is less discordant, and thus an upper-intercept age was not calculated.

Patton and Dutro (1969) correlated the Devonian–Triassic carbonate succession on St. Lawrence Island with Devonian through Triassic rocks of the Brooks Range. Lithofacies and biofacies in the Devonian–Triassic carbonate succession on St. Lawrence Island match well with those of the Devonian through Triassic units of the Kelly River allochthon in the western and central Brooks Range (Figs. 3, 4, 5, and 7). Devonian sections in both successions are thick, dominated by dolostone, and were deposited largely in shallow-water to very shallow-water settings (Figs. 4, 5, and 7). Both Mississippian sections consist of fossiliferous limestone and dolostone, with lower intervals that contain appreciable interbeds of quartzose sandstone, and chert nodule–bearing upper intervals that deepen upward. Both successions are capped by thin, chert-rich intervals. The lower part of the capping interval in the Kelly River allochthon is the Siksikpuk Formation, which may correlate with the undated cherty unit on St. Lawrence Island. The highest beds in both successions are radiolarian- and bivalve-bearing chert and limestone of Middle through Late Triassic age (Otuk Formation in the Kelly River allochthon). Numerous specific microfacies throughout both successions are essentially identical, including Devonian peloidal dolostone, Mississippian quartz-carbonate sandstone and bioclastic grainstone, and Upper Triassic monotid limestone with calcitized radiolarians (Figs. 4, 5, and 7).

On the basis of petrography, the most probable sources of compositionally mature siliciclastic detritus in the Ml strata on St. Lawrence Island and the Mississippian Utukok Formation in the western Brooks Range were sedimentary rocks that were deposited prior to development of the pre-Mississippian unconformity. The development of the unconformity and long duration of exposure in the Ellesmerian highlands allowed these pre-Mississippian strata to be reworked first into rift basins (Endicott Group) and later into passive-margin settings of the Ellesmerian megasequence (Moore et al., 1994). Petrographically, Ellesmerian megasequence sandstones and conglomerates are typically well rounded and compositionally supermature, consisting predominantly of quartz grains, chert, and quartzite lithics. The texture and composition of these Mississippian-age rocks attest to the important role of chemical weathering and mechanical reworking prior to their deposition. The quartz arenite sandstone lenses within the western Brooks Range Utukok Formation and St. Lawrence Island Ml strata are no exception, and they occur as minor components of carbonate-dominant successions deposited in carbonate bank settings.

The combined DZ ages from the three samples of the Ml unit of the Devonian–Triassic carbonate succession from St. Lawrence Island (Fig. 4) form a distinctive age distribution with three prominent modes between ca. 700 and 400 Ma, a mode between ca. 1000 and 900 Ma, and abundant zircons in the age range of ca. 1900–1100 Ma (Figs. 9A–9D). The Mississippian Utukok Formation from the Kelly River allochthon in the Brooks Range (Fig. 7) has a similar overall DZ age distribution, including the three young (ca. 700–400 Ma) modes, a gap ca. 900–700 Ma, and several Proterozoic modes between ca. 1900 and 950 Ma (Fig. 9E). Numerous localities from which similarly distinctive age distributions have been reported are spread along the Arctic Alaska–Chukotka microplate Arctic margin for over 100° of longitude (Fig. 1A). Other samples from the Arctic Alaska–Chukotka microplate that have DZ age distributions similar to that of the St. Lawrence Ml unit come from the Kapaloak sequence on Lisburne Peninsula (Figs. 1B and 12; Miller et al., 2010), a heterolithic Lisburne Group turbidite sample of Mississippian age from the Red Dog mine area in the western Brooks Range (Figs. 1B and 12; Ikalukrok unit; Dumoulin et al., 2014b; although that sample has a larger proportion [~20%] of ages younger than 400 Ma), Carboniferous clastic rocks found within the Wrangel Island carbonate sequence (Figs. 1B and 12; Miller et al., 2010), and Carboniferous strata as far west as the New Siberian Islands (Figs. 1A and 12; Kotelny and Belkovsky Islands; Ershova et al., 2015a, 2015b).

Despite the relative ubiquity of this age distribution in Arctic Alaska–Chukotka microplate Carboniferous rocks (Fig. 12), the same age distribution appears to be absent or diluted in other Mississippian rocks elsewhere in Arctic Alaska. For example, published DZ ages from three samples of the Mississippian Kekiktuk Conglomerate have considerably different zircon age distributions than those in our study (Fig 12; Inigok #1 North Slope sample and two Mt. Doonerak area samples combined in Fig. 12; Gottlieb et al., 2014; Strauss et al., 2017). Notably, these Kekiktuk Conglomerate samples contained substantial amounts of Devonian zircons, representing derivation from a source region that contrasts with that of the St. Lawrence Island Ml DZ samples, which only have a few Early Devonian ages in their distribution (Fig. 12). The predominance of Devonian age zircons in Kekiktuk samples may simply be due to a basin-scale effect on sediment sampling (e.g., Gottlieb et al., 2014, and references therein) with Kekiktuk strata having been more “point-sourced” (i.e., higher order and therefore less well-mixed) than Ml samples within the overall basin setting. Kekiktuk Conglomerate strata may have also been deposited in areas where locally exposed Devonian-age plutons were substantially more fertile zircon sources than other erodible zircon sources, thus effectively amplifying the Devonian age mode in the overall age distribution (Fig. 12).

The paucity of Late Devonian zircons in St. Lawrence Island Ml and western Brooks Range Kelly River allochthon Utukok Formation samples (Fig. 9F) suggests that unroofed Devonian plutons in the Arctic Alaska basement (age range mostly 395–360 Ma; e.g., Amato et al., 2014; Ward et al., 2019) were not important sources and that the Ml and Kelly River allochthon Utukok strata were derived from different catchment areas than those feeding the Kekiktuk Conglomerate strata (Fig. 12). Given that Mississippian Kekiktuk Conglomerate strata, with more prominent Devonian zircon ages (Fig. 12), are located farther northeast relative to St. Lawrence Island and the western Brooks Range Kelly River allochthon Utukok Formation, the magnitude of late Paleozoic stratigraphic unroofing may have been lower in western than in eastern (present day reference frame) parts of Arctic Alaska, inhibiting exhumation of any Devonian plutons that were near the Ml and Utukok strata during their deposition.

In the Canadian Arctic Islands, a relevant tectonic example of Devonian DZ source unroofing is evident in the Lower Mississippian Borup Fiord Formation of Ellesmere Island (Beauchamp et al., 2019). The Borup Fiord Formation (including previously misidentified Okse Bay Formation strata rectified by Beauchamp et al., 2019) shows dramatic increases in the modal abundance of Devonian-age zircons up section, whereas the rest of the overall age distribution is generally static (fig. 9 inBeauchamp et al., 2019). The Borup Fiord Formation was deposited during a period of active rifting, and the sudden influx of Devonian-age DZ as one moves up section was explained by Beauchamp et al. (2019) as a result of rift-related exhumation and unroofing of Devonian plutons. Borup Fiord Formation strata were interpreted as having been deposited in a reasonably well-integrated drainage system that had DZ sources from Silurian- and Devonian-age synorogenic strata (Beauchamp et al., 2019), and thus they may be a good analogy for similar mixing processes that affected the clastic sediments of the Ml and Kelly River allochthon Utukok Formation samples, although the depositional environments of the clastic-dominated Borup Fiord Formation and carbonate-dominated Ml and Kelly River allochthon Utukok samples are clearly not analogous.

The amalgamation of KDE modes between ca. 640 Ma and ca. 440 Ma in the St. Lawrence Island Ml samples and in the other Carboniferous samples with correlative age distributions suggests their sources shared a common characteristic of having a mix of zircons of these ages (Fig. 11). These late Neoproterozoic to Silurian ages are consistent with derivation from magmatic rocks associated with the Caledonian and Timanian orogenies, or potentially other paleogeographic regions with similar age sources (e.g., Arctic Alaska–Chukotka microplate basement for Neoproterozoic-Cambrian modes and oceanic arcs for the Silurian mode are alternative possibilities). Similar age modes in the ca. 700–400 Ma range have been documented in pre-Mississippian Paleozoic strata in the circum-Arctic region (e.g., Lorenz et al., 2008; Amato et al., 2009; Miller et al., 2011; Anfinson et al., 2012; Hadlari et al., 2014; Beranek et al., 2015; Johnson et al., 2016; Strauss et al., 2017, 2019; Ershova et al., 2018a, 2018b; Malone et al., 2019; Robinson et al., 2019). Thus, it seems possible that at least parts of the distinctive age distribution seen in Carboniferous strata were derived from erosion of pre-Mississippian strata that were exposed in the Arctic Alaska–Chukotka microplate Ellesmerian highlands (sensu Homza et al., 2020).

Another possible source for the 640–440 Ma DZ ages could have been Endicott Group strata (e.g., Kanayut Conglomerate and Noatak Sandstone; Nilsen and Moore, 1984; Moore and O'Sullivan, 2019), which were derived from the Ellesmerian highlands in Late Devonian–Early Mississippian time, but in part predate Ml and Utukok Formation strata (e.g., Dumoulin et al., 2004, 2006). This may have occurred if unconsolidated (i.e., surficial) and long-weathered deposits of Endicott Group strata that existed landward of the Ellesmerian margin were transgressed by the marine environment during Mississippian time, reworked by nearshore processes that contributed to efficient mixing, and redeposited in the carbonate-dominated Ml and Utukok carbonate-dominated environments. Furthermore, if St. Lawrence Island Ml and western Brooks Range Kelly River allochthon Utukok Formation strata were indeed sourced from Endicott Group strata (as opposed to older strata beneath the pre-Mississippian unconformity), the same logic regarding sediment sourcing from exposed pre-Mississippian rocks also seems applicable for DZ sources of the Endicott Group. Later reworking into the carbonate-dominated Ml and Utukok strata would have then been a simple case of sediment recycling—a hypothesis that is likely testable in the western Brooks Range by determining DZ age distributions in Endicott Group strata that were onlapped after deposition of Utukok Formation strata. In summary, the most likely sources of well-mixed, compositionally mature siliciclastic material in the Ml strata on St. Lawrence Island and in Mississippian Utukok Formation strata in the western Brooks Range were rocks deposited prior to development of the pre-Mississippian unconformity and exposed in the Ellesmerian highlands, or Endicott Group strata that were derived from those same pre-Mississippian rocks.

The distinctive lithology and unimodal DZ age distributions of the feldspathic subunit of Mississippian(?)–Triassic clastic succession samples are strikingly similar to those of the Nuka Formation in the Nuka Ridge allochthon (Fig. 9). Both units consist of arkosic sedimentary strata of late Paleozoic age containing DZs of almost exclusively middle Paleoproterozoic age (ca. 2.06–2.03 Ga). Although DZ age distributions from the Nuka Formation are slightly broader than those from the St. Lawrence Island samples, and calculated MDAs for the Nuka Formation are slightly different (Table 1), the combination of mid-Paleoproterozoic DZs and late Paleozoic fossils in both units is striking and has not been reported from elsewhere in the Arctic.

Potential sources of mid-Paleoproterozoic DZs were discussed by Moore et al. (1997) and Bradley et al. (2014). Paleoproterozoic crust has not been reported in the Arctic Alaska–Chukotka microplate, but several Paleoproterozoic outcrop localities are known in west-central Alaska (see review in Bradley et al., 2014). Igneous rocks of the Kilbuck terrane (Fig. 1A; Jones et al., 1987) yielded an age of ca. 2.05 Ga (Turner et al., 1983), and rocks in the Idono Complex yielded an upper-intercept U-Pb zircon age of ca. 2.06 Ga (age locality 12 in Fig. 1B; Miller et al., 1991). These are granitic rocks with no clear connection to surrounding rock assemblages, which are significantly younger, consistent with the assembly of Alaska from disparate terranes (e.g., Silberling et al., 1994). The distinctively old age of ca. 2.0 Ga basement rocks in the Idono Complex and Kilbuck terrane, and the presence of DZs of this age in the Farewell terrane and Brooks Range, led Bradley et al. (2014) to postulate connections between these areas that are now widely separated across Alaska. However, at least parts of the Kilbuck terrane apparently underwent amphibolite-to eclogite-grade metamorphism ca. 1.77 Ga, according to an abstract by Turner et al. (1983). If taken at face value, these results are incompatible with the thermal history of the Nuka Formation source region based on thermochronology of detrital rutile and microcline from the Nuka Formation, which indicates maximum temperatures below 400 °C after ca. 1.9 Ga (Moore et al., 1997). These data suggest that the high-grade crust in the Kilbuck terrane itself was not the source of the DZs found in the Nuka Formation, but instead the source was a distinct but similar-age crustal fragment. Additionally, if modern thermochronologic data were obtained from other parts of Kilbuck terrane, it might allow for the possibility that the Kilbuck, Idono, or other presently unknown Proterozoic blocks may have been sources for the Paleoproterozoic zircons in the Nuka Formation.

The DZ age distributions in the feldspathic subunit of Mississippian(?)–Triassic clastic succession on St. Lawrence Island and the Nuka Formation in the Nuka Ridge allochthon of the western Brooks Range suggest an important change in provenance. The unimodal age distribution in these strata is not evident in the other Paleozoic samples from this study, nor has it been identified in samples from published studies elsewhere in the Arctic Alaska–Chukotka microplate (e.g., Amato et al., 2009). This suggests that Nuka Formation detritus was deposited in a geographically limited part of the Arctic Alaska–Chukotka microplate, perhaps due to removal of the Nuka crustal source by rifting (e.g., Moore et al., 1997), or due to a relatively short-lived tectonic juxtaposition of that source terrane with the outboard edge of the Arctic Alaska–Chukotka microplate, consistent with the original interpretation of Tailleur and Sable (1963, p. 632), which stated that “large-scale structural dislocations” were involved. The source for the Kilbuck terrane rocks and for the Nuka Formation may have been a tectonically dispersed continental fragment of Siberia (e.g., Donskaya, 2020) or another craton left behind during Proterozoic or Devonian rifting of the Arctic Alaska margin (Moore et al., 1997; Ershova et al., 2020). The juxtaposition of this fragment against the distal edge of Arctic Alaska is consistent with the model of McClelland et al. (2021), invoking long-term transform motion of terranes around the Arctic if it occurred during the Devonian or later. The timing of deposition of the feldspathic subunit of Mississippian(?)–Triassic clastic succession and that of the Nuka Formation in the western Brooks Range are consistent with transpressional/transtensional deformation interpreted in the subsurface of northern Arctic Alaska (e.g., Hubbard et al., 1987; Fulk, 2010).

The DZ age distributions in the Triassic graywacke subunit of Mississippian(?)–Triassic clastic succession on St. Lawrence Island are similar to age distributions reported from other Triassic strata in the Arctic, including the Lisburne Peninsula (Miller et al., 2006; Dumoulin et al., 2018, 2022b), Chukotka (Miller et al., 2006, 2010; Amato et al., 2015), Wrangel Island (Miller et al., 2010), the Verkhoyansk belt (Miller et al., 2013), and the northern Sverdrup Basin (Miller et al., 2006; Omma et al., 2011; Alonso-Torres et al., 2018; see Fig. 13). However, the composition, depositional setting, and DZ age distributions in the graywacke subunit of Mississippian(?)–Triassic clastic succession are distinctly different from those of Triassic quartz-rich sandstones found in shelfal successions in the North Slope subsurface of Alaska (Gottlieb et al., 2014), in the northeastern Brooks Range (Gottlieb et al., 2014), and in the southern Sverdrup Basin of the Canadian Arctic Islands (Miller et al., 2006; Midwinter et al., 2016), all of which were deposited in shallow-marine settings and have DZ age distributions that are exclusively Devonian and older and mostly older than 1.0 Ga (Fig. 13). Detrital zircon age distributions similar to those from the graywacke subunit of the Mississippian(?)–Triassic clastic succession samples include those from Triassic turbidites interbedded with the uppermost part of the Otuk Formation on Cape Lisburne (Lisburne Peninsula) in northwestern Alaska (Fig. 1B; Miller et al., 2006; Dumoulin et al., 2018, 2022b), Triassic turbidite successions on Wrangel Island and in the Chukotka area of eastern Russia (Fig. 1B; Miller et al., 2006, 2010; Amato et al., 2015), and Triassic shallow-marine sandstones in the northern Sverdrup Basin (Fig. 1A; Omma et al., 2011; Midwinter et al., 2016; Fig. 13). Some of these strata, in particular, those from the Lisburne Peninsula and most of the Wrangel Island and Chukotka rocks, are micaceous, heterolithic graywackes generally similar in composition to the graywacke subunit of the Mississippian(?)–Triassic clastic succession, although a few of the Wrangel Island and Chukotka samples consist mainly of quartz grains. Miller et al. (2006, 2010, 2013), Gottlieb et al. (2014), Anfinson et al. (2016), and Midwinter et al. (2016) have suggested sediment provenance to the present-day northwest for Triassic strata that contain age distributions with Triassic ages, and the graywacke subunit on St. Lawrence Island may also have been derived from such sources (“Northwestern sources” on Fig. 13).

The samples that we dated from St. Lawrence Island that produced Triassic MDAs (70APa-103, 71APa-193, 71APa-194) have modal abundances of ages that match biostratigraphic age constraints. Similarly young ages have also been observed in numerous samples from Chukotka, the Lisburne Peninsula, and other circum-Arctic areas (Fig. 13). These Triassic zircons may have been derived from Eurasian source regions (Miller et al., 2013; Anfinson et al., 2016) and/or a volcanic complex that fringed the northern Laurentian margin of Pangea (Hadlari et al., 2018). By contrast, Triassic–Jurassic strata from the North Slope and northeast Brooks Range of Arctic Alaska and from the southern Sverdrup Basin must have been depositionally isolated from sources of Triassic zircons because they lack young zircons, and they appear to be recycled from older strata deposited along the northern Laurentian margin (e.g., Gottlieb et al., 2014; Midwinter et al., 2016; Fig. 13). Our data thus place St. Lawrence Island during Triassic time in a position near Chukotka, the Lisburne Peninsula, and the northern Sverdrup Basin.

In a comprehensive analysis of sources for Permian–Triassic DZ grains in the Sverdrup Basin in the Canadian Arctic Islands, Anfinson et al. (2016) suggested that the Taimyr region of the Uralian orogen was the source. Alternatively, Hadlari et al. (2018) suggested that there was an arc to the northwest of the Sverdrup Basin supplying zircons of late Permian to Triassic age. These two hypotheses for the source of the Triassic zircons in other Arctic regions, such as on St. Lawrence Island, could be tested by analyzing Hf and O isotopic ratios, trace-element concentrations, and morphology of Permian–Triassic zircons.

The DZ data partly support the correlation made by Patton et al. (2011) between the graywacke subunit of the Mississippian(?)–Triassic clastic succession on St. Lawrence Island and rocks of Chukotka (Fig. 1B), but, given the lack of DZ data from late Paleozoic age strata in Chukotka, it is impossible to extend this correlation earlier than the Triassic intervals of each area. Likewise, the correlation made by Patton et al. (2011) between the Angayucham terrane in western Alaska (Fig. 1B) and the Tomname terrane (i.e., the composite section that includes both the feldspathic and graywacke subunits of the Mississippian[?]–Triassic clastic succession and the spatially associated Permian–Triassic gabbro and diabase TrPg unit) cannot be assessed using DZ data, because no DZ analyses have been published from Angayucham terrane strata. However, this correlation seems less likely based purely on lithologic and stratigraphic criteria, as siltstone and sandstone beds are not reported to be a component of the Angayucham terrane (e.g., Young, 2004), whereas they make up as much as a quarter of the graywacke subunits of the Mississippian(?)–Triassic clastic succession on St. Lawrence Island.

The Tozitna terrane (Fig. 1B; Silberling et al., 1994; Dover, 1994), which has Mississippian–Triassic sedimentary rocks and coeval mafic igneous rocks, has been correlated with the Angayucham terrane based on structural and lithologic relationships (Patton et al., 1994) and may better correlate with St. Lawrence Island stratigraphy. However, a DZ sample from the Tozitna terrane has a Devonian DZ age mode (Bradley et al., 2007) not observed in St. Lawrence Island samples. Additional DZ analyses from sedimentary rocks and geochemical comparisons between the mafic igneous rocks on St. Lawrence Island and those in the Angayucham-Tozitna terranes would further test these proposed correlations.

The relationship of St. Lawrence Island to the Arctic Alaska–Chukotka microplate has been controversial. The terrane map compilation of Silberling et al. (1994) did not include St. Lawrence Island in the Arctic Alaska–Chukotka microplate. Amato et al. (2003) suggested that the presence of Angayucham-type mafic rocks on the island meant that the southern boundary of the Arctic Alaska–Chukotka microplate ran through the island. The controversy stems from the lack of Proterozoic basement rocks on St. Lawrence Island that can be correlated to basement in the Brooks Range, Seward Peninsula, and Wrangel Island. Dating of xenoliths from Quaternary basalt on St. Lawrence Island (Fig. 2A) yielded only Cretaceous and younger ages (Miller et al., 2002). Our study confirms correlations between late Paleozoic–Triassic strata on St. Lawrence Island and those in the Brooks Range allochthons, the Lisburne Peninsula, and Wrangel Island, and it provides strong support for the conclusion that the rocks we studied on St. Lawrence Island constitute an important part of the Arctic Alaska–Chukotka microplate (Fig. 2A).

Despite its proximity to Seward Peninsula (Fig. 1B), the geology of St. Lawrence Island is strikingly different. Most of the pre-Quaternary exposures on Seward Peninsula are metamorphosed igneous rocks ranging in age from ca. 550 Ma to ca. 970 Ma (Amato et al., 2014) and metasedimentary rocks that generally have MDAs of late Neoproterozoic/Cambrian to Devonian (one sample had an MDA of Pennsylvanian; Amato et al., 2009; Till et al., 2014). Strata on the northern part of Seward Peninsula (York Mountains and Teller regions) preserve sedimentary structure but are generally older than Mississippian (Sainsbury, 1969; Dumoulin et al., 2014a). Thus, no clear stratigraphic correlations can be made between Seward Peninsula and St. Lawrence Island, despite their current proximity, other than the presence of latest Neoproterozoic DZ ages in samples from both Seward Peninsula and St. Lawrence Island that can be matched with ages of igneous rocks on Seward Peninsula and across the Bering Strait in Chukotka (Amato et al., 2009; Gottlieb et al., 2018).

The modern position of St. Lawrence Island, isolated from mainland Alaska, is likely the result of a combination of (1) southward displacement from the area along strike of the Brooks Range as a result of Cretaceous extension documented in the southern Brooks Range (Miller and Hudson, 1991) and on Seward Peninsula (Amato et al., 1994; Dumitru et al., 1995; Amato and Miller, 2004); and (2) Cenozoic extension in the Bering Straits region (e.g., Worrall, 1991). This could have formed the salient (Amato et al., 2004), originally interpreted as an “oroclinal bend” (Patton and Tailleur, 1977), in the map pattern of the southern margin of the Arctic Alaska–Chukotka microplate where it crosses from the southern Brooks Range into Chukotka (Fig. 1B). This follows the trend of Triassic–Jurassic mafic rocks that appears in the Angayucham terrane and on St. Lawrence Island (map unit TrPg; Figs. 2A and 2B).

The displacement of St. Lawrence Island from these other Arctic Alaska–Chukotka microplate constituents could also have occurred along a strike-slip fault. Notably, near the eastern boundary of Seward Peninsula with the rest of Alaska, there is the Kugruk fault zone (e.g., Till et al., 1986). Plafker and Berg (1994) suggested that this was a strike-slip fault, and Amato et al. (2004) suggested that a left-lateral fault could have displaced the Seward Peninsula to the south relative to the Brooks Range and the rest of mainland Alaska. Metamorphic rocks of the Nome Group are mapped on both sides of this boundary, so the actual structure separating Seward Peninsula from the Yukon-Koyukuk Basin is unexposed, but it may be parallel to the Kugruk fault zone. Toro and Amato (2015) reported shear-sense indicators from what they referred to as the Kugruk shear zone that suggested right-lateral displacement possibly related to opening of the Canada Basin (Fig. 1), but it is unclear if these indicators represent the total displacement or just the latest phase. Our study is more consistent with earlier hypotheses of left-lateral offset of St. Lawrence Island relative to the correlative areas of the Brooks Range and the Lisburne Peninsula and supports models of widespread translation in the Bering Straits region (e.g., McClelland et al., 2021). This study strengthens links between the Brooks Range and St. Lawrence Island but does not directly test the validity of rotational models for the Cretaceous opening of the Canada Basin.

Our data indicate the presence on St. Lawrence Island of upper Paleozoic rocks equivalent to upper allochthons exposed in the western and central Brooks Range, as well as Triassic turbidites likely derived from the present-day northwest based on provenance data presented here and in previous studies. Lithofacies and fossils of the Devonian–Triassic carbonate succession on St. Lawrence Island match well with those of the Kelly River allochthon in the western and central Brooks Range, and DZ spectra from Mississippian quartz arenites in both areas are similar. These data require that St. Lawrence Island in the late Paleozoic was in a relatively distal position on the Arctic Alaska continental margin, adjacent to the Angayucham ocean basin (Fig. 3A).

The DZ populations in the Ml unit on St. Lawrence Island and in the Utukok Formation of the Kelly River allochthon were likely recycled from older updip strata. The DZ age distributions in these samples have multiple Ordovician–Silurian and latest Neoproterozoic–Cambrian modes also found in many Carboniferous-age strata across the Arctic Alaska–Chukotka microplate. The previous host strata of the DZs in the Ml and Utukok Formation strata were likely pre-Mississippian rocks in the Ellesmerian highlands of Arctic Alaska and/or the Chukchi Platform. It is plausible that some of these zircons were first weathered out of pre-Mississippian rocks and deposited as Late Devonian–Early Mississippian Endicott Group strata proximal to the eroding highlands, and then these unconsolidated deposits were recycled later in Mississippian time as clastic pulses into more distal carbonate-dominated environments.

The Nuka Formation of the Nuka Ridge allochthon in the Brooks Range is a striking match in its petrography and DZ age distributions with the feldspathic subunit of the Mississippian(?)–Triassic clastic succession on St. Lawrence Island. The source for the zircons in these units was likely a Paleoproterozoic crustal fragment that was rifted from and/or translated along the continental margin (e.g., Moore et al., 1997), shedding sediments with 2.06 Ga zircons along Arctic Alaska's distal margin (Fig. 3A). This age matches that of the Kilbuck terrane in southwestern Alaska (e.g., Bradley et al., 2014), although a source-sink relationship between the Kilbuck terrane and the feldspathic subunit of the Mississippian(?)–Triassic clastic succession on St. Lawrence Island (and correlative Nuka Formation) is problematic because of apparent disparate cooling histories of minerals in Kilbuck crust and Nuka Formation strata.

Lithic-rich turbidites in the graywacke subunit of the Mississippian(?)–Triassic clastic succession on St. Lawrence Island have DZ spectra that resemble those of Triassic turbidites interbedded with the Otuk Formation on Cape Lisburne in northwestern Alaska and Triassic turbidites exposed on Wrangel Island and in Chukotka in northeastern Russia, as well as those from Triassic sandstones in the northern Sverdrup Basin of Canada. Sources to the present-day northwest, such as the Chukchi Platform, Taimyr, and/or the Uralian orogen (e.g., Gottlieb et al., 2014; Anfinson et al., 2016; Midwinter et al., 2016), may have produced the Carboniferous through Triassic zircons that characterize all of these units. Alternately, the Triassic zircons may have been sourced from an arc fringing the northern Laurentian margin of Pangea (e.g., Hadlari et al., 2018); our data do not directly address this controversy. The data presented in this contribution allow for new tests of relationships with other terranes (Fig. 1) that are associated with the late Paleozoic–early Mesozoic evolution of the Amerasian Arctic region.

The Energy and Mineral Resources Program of the U.S. Geological Survey provided funding. Andy Allard, Beth Drewes, and Jill Schneider made access to archival materials and sample retrieval possible, and Alan Pongratz and Nora Shew assisted with map compilation. Special thanks go to Bill Patton, Bela Csejtey, Gil Mull, and their coworkers for inspiring this study and collecting the materials from which it was built. The authors appreciate conversations with Elizabeth Miller and Jaime Toro. Maurice Colpron provided assistance with Figure 1A. Paul O'Sullivan at GeoSep Services separated and analyzed three of the detrital zircon samples from the Brooks Range and graciously contributed multiple follow-up analyses of Nuka sample DT84–14/15. The remaining detrital zircon ages were acquired at the Arizona LaserChron Center, supported by National Science Foundation grant EAR-1649254, with the help of George Gehrels, Mark Pecha, and Dominique Giesler. Detailed comments from U.S. Geological Survey reviewers Steve Box and Jamey Jones, two anonymous reviewers for Geosphere, and Associate Editor Todd LaMaskin are greatly appreciated. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. government.