First bedrock samples dredged from submarine outcrops in the Chukchi Borderland, Arctic Ocean

The Arctic Ocean consists of two subbasins, the Eurasia and Amerasia Basins (Fig. 1). The Eurasia Basin is floored by oceanic crust and is the northernmost continuation of the mid-Atlantic Rift (Heezen and Ewing, 1961). Magnetic anomalies within the Eurasia Basin indicate that seafloor spreading began at ca. 60 Ma, separating the Lomonosov Ridge from the Barents Shelf (Ostenso and Wold, 1973; Herron et al., 1974). In contrast, the crustal character and plate tectonic history of the adjacent Amerasia Basin are very poorly understood. The Amerasia Basin includes a number of intrabasinal bathymetric highs including the Chukchi Borderland that appear to represent continental crust (Fig. 1). The nature and geologic history of the bathymetric features in the Arctic Ocean have implications for both natural resource exploration and for Extended Continental Shelf (ECS) submissions of Arctic states under Article 76 of the United Nations Convention on the Law of the Sea (UNCLOS). The geologic make up of bathymetric features is important in defining the “natural prolongation” of land territory for the purpose of defining an Extended Continental Shelf under UNCLOS. The final and binding international boundaries that result from these submissions will have far-reaching implications for future marine science research in the Arctic Ocean and consequences for scientific studies monitoring and exploring this important region.

Few geologic samples have been recovered from the Arctic Ocean, relegating the study of its tectonic and rift history to the surrounding landmasses. The only drill core available in the Arctic Ocean is from the Lomonosov Ridge (Moran et al., 2006). Prior to recent data collection by Arctic states for their ECS projects, rock samples within the Amerasia Basin were limited to one successful dredge of an outcrop on the Alpha Ridge (Forsyth et al., 1986; Van Wagoner et al., 1986) and small rock chips from sediment cores (Grantz et al., 1998) on Northwind Ridge.

In order to answer first-order questions about the geologic nature of the intrabasinal features, dredged rock samples from submarine outcrops within the Amerasia Basin were collected aboard the icebreaker U.S. Coast Guard Cutter (USCGC) Healy between 2008 and 2012 for the U.S. ECS project. One of these dredges (HLY0905-DS5; Mayer and Armstrong, 2009) produced hundreds of kilograms of meta-igneous rocks sampled from a steep slope in the central Chukchi Borderland (Figs. 1 and 2). The aim of this paper is to describe the nature of the basement underlying the Chukchi Borderland and to evaluate possible geologic relationships of the Borderland to the Arctic region. Below we report petrologic, zircon U-Pb age, whole-rock compositions, and radiogenic isotope data for the major lithologies recovered from dredge HLY0905-DS5.

The Chukchi Borderland is a bathymetric high of presumed continental affinity that is physiographically connected to the Alaska margin (Fig. 1; e.g., Hall, 1990). The borderland covers ∼420,000 km² and is deformed and extended by major north-south–trending normal faults (Hall, 1990; Brumley, 2009). The geologic and tectonic history of the Chukchi Borderland has been controversial. Grantz et al. (1998) sampled talus-slope fragments from sites along the Northwind Ridge in the Chukchi Borderland (red X, Fig. 1) using sediment piston cores and box cores. The lithologies represented within the sediment core samples were interpreted to resemble the Franklinian Neoproterozoic–early Paleozoic passive margin sequence of northern Canada (described below). Grantz et al. (1998) concluded that the rock samples appeared to correlate most closely to shelf strata situated between the north end of the Mackenzie Delta and Melville Island with an assumed attachment point of the Chukchi Borderland to the Banks Island area (striped region, Fig. 1). Based upon preliminary analysis of the HLY0905 dredge results, Brumley et al. (2013) alternatively proposed that the Chukchi Borderland was related to peri-Laurentian orogenic terranes similar to southwestern Svalbard and the Pearya terrane of northern Ellesmere Island.

The Grenville orogeny affected northern Laurentia during assembly of the supercontinent Rodinia from 1200 to 950 Ma (Davidson, 2008). Evidence for Grenville age deformation, metamorphism, and magmatism occurs in northeastern Greenland (e.g., Higgins and Leslie, 2008; Kalsbeek et al., 2008), the assembled terranes of Svalbard (e.g., Johansson et al., 2005), and the Pearya terrane of northern Ellesmere Island (e.g., Trettin, 1987) leading Johansson et al. (2005), to argue for a northern arm of the Grenville orogen in the Arctic (Fig. 1). Amalgamation of the composite terrane of Pearya is attributed to collision of a continental fragment with an island arc during the Cambro-Ordovician M’Clintock orogeny. This orogenic event involved thrusting of calc-alkaline volcanic rocks, deep marine sediments, and a dismembered ophiolite over the Proterozoic basement of Pearya (Trettin 1987, 1991; Trettin et al., 1992). M’Clintock orogenesis is coeval with subduction in southwestern Svalbard (Eidembreen event), where high-pressure metamorphic rocks (blueschists and eclogites) formed during subduction-related collision in Cambro-Ordovician time (Ohta et al., 1989; Gee and Page, 1994; Harland, 1997; Gee and Teben’kov, 2004; Gee et al., 2008; Labrousse et al., 2008). Closure of the early Paleozoic Iapetus Ocean led to the collision of eastern Laurentia (North America) and Greater Baltica (Eurasia) resulting in the Silurian to Early Devonian Caledonian orogeny (e.g., Leslie and Higgins, 2008). The Caledonian orogen, which straddles both conjugate margins of the Atlantic, continues into the high Arctic, where it is truncated at the shelf edges of the modern Arctic Ocean (Fig. 1). The Scandian phase of the Caledonian orogeny affected northeastern Greenland, Norway, and parts of Svalbard between ca. 435 and 415 Ma (Fig. 1; Harland, 1997; McKerrow et al., 2000; Johansson et al., 2005; Gee et al., 2008; Leslie and Higgins, 2008). The Ellesmerian orogeny created a south-verging fold belt that affected Arctic Alaska, Pearya, northern Greenland, and parts of western Svalbard (Fig. 1; e.g., Trettin and Norford, 1994). This deformation is thought to reflect collision of an unknown landmass with the northern margin of Laurentia from Devonian to earliest Carboniferous time (Embry, 1993; Trettin and Norford, 1994; Colpron and Nelson, 2006). This landmass was subsequently rifted away by the opening of the Amerasia Basin and may exist as rifted fragments around the Arctic shelf margins and within the Arctic Ocean itself (Beranek et al., 2010).

The Franklinian margin of northern Laurentia consisted of an uninterrupted latest Neoproterozoic–Devonian continental platform that formed along the Arctic edge of Laurentia and created associated slope and deep-water basin deposits (Trettin, 1991; Henricksen and Higgins, 1998). On the platform, carbonate deposition predominated over this lengthy period, constructing a thick pericratonic wedge (Harrison et al., 1991). Lower Paleozoic sedimentation on the Arctic platform and within the Franklinian basin continued until Devonian uplift, deformation, and foreland subsidence associated with the Ellesmerian orogeny (Haimila et al., 1990; discussed below).

The bedrock geology exposed on Banks Island in the Canadian Arctic Islands, to which the Chukchi Borderland has been correlated (Grantz et al., 1998), includes undeformed Meso- and Neoproterozoic dolostones and sandstones (Trettin, 1989) with rift-related diabase sills and extrusive basaltic rocks (ca. 723 Ma; Heaman et al., 1992). Cambrian to Devonian strata of the Arctic platform and Franklinian basin are overlain by shallow marine and fluvial deltaic clastic sediments of the Ellesmerian foreland basin (Trettin, 1991; Hadlari et al., 2012).

The crystalline basement of the Pearya terrane (Fig. 1) consists of Mesoproterozoic metavolcanics, orthogneisses, and metasedimentary rocks that were affected by late Mesoproterozoic to early Neoproterozoic (ca. 1100–960 Ma) deformation and magmatism attributed to Grenville orogenesis as expressed in the North Atlantic region (Succession I of Trettin, 1987). These rocks are overlain by Neoproterozoic to Ordovician age platform strata and volcanic rocks that are interpreted to have been deposited in an extensional tectonic setting (Succession II of Trettin, 1987). Additional mafic volcanic and sedimentary rocks occur adjacent to the platform sequence and represent deposits that formed during collision of the continental platform with an island arc during the Cambro-Ordovician M’Clintock orogeny (Succession III of Trettin, 1987). The subduction-related rocks associated with this collision are unconformably overlain by Middle Ordovician calc-alkaline volcanic rocks (Succession IV of Trettin, 1987, 1991) and Late Ordovician to Silurian clastic rocks that record continuous subsidence without volcanism or deformation (Trettin, 1987).

Pearya was adjacent to the northern edge of Laurentia by Late Silurian time as demonstrated by clastic rocks that record proximity to the Laurentian margin and subsequent Late Silurian to Early Carboniferous accretion-related deformation (Succession V of Trettin, 1991; Churkin and Trexler, 1980; Trettin, 1987). The transpressional accretion of Pearya resulted in the development of the Clements-Markham fold belt on northern Ellesmere Island (Fig. 1; Piepjohn et al., 2007).

The Pearya Terrane has been correlated to the southwestern terranes of Svalbard (e.g., Trettin, 1987; Harland, 1997). Like Pearya, southwestern Svalbard consists of Proterozoic basement with deformation, metamorphism, and anatectic magmatism between 1160 and 940 Ma associated with the Grenville orogeny (Ohta et al., 1989; Harland, 1997; Johansson et al., 2005). Parts of southwestern Svalbard experienced metamorphism and magmatism between ca. 650 and 620, and ca. 620–540 Ma (Peucat et al., 1989; Majka et al., 2008, 2012). Ediacaran tillites record rift-associated uplift and glaciation (Harland, 1997). Cambro-Ordovician lithotectonic units along the western coast of Svalbard include a subduction complex with high-pressure metamorphic and oceanic volcanic rocks that are thought to represent a collision of an island arc along the northern Laurentian margin (Ohta et al., 1989). The subduction complex is unconformably overlain by Late Ordovician through Early Silurian island arc and shelf sedimentary sequences (Ohta et al., 1989; Ohta, 1994; Harland, 1997; Mazur et al., 2009).

Neither Pearya nor southwestern Svalbard record evidence for Silurian (Scandian) Caledonian metamorphism (Gasser and Andresen, 2013). The translation and accretion of western Svalbard and Pearya along the northern edge of Laurentia is coeval with late Caledonian orogenesis along the northeastern edge of Laurentia and was probably completed by Early Devonian time (Trettin, 1987; von Gosen et al., 2012).

Deformed and metamorphosed amphibolite-grade igneous rocks including biotite-bearing amphibolites, garnet-bearing feldspathic gneisses, and potassium feldspar augen-orthogneisses were dredged from a steep normal fault scarp in central Chukchi Borderland. When dredging in ice-covered water, it is essential to maximize the chance that dredged rock samples are from submarine bedrock exposures by avoiding locations dominated by ice-rafted debris (IRD). To avoid collecting the thick pelagic mud and IRD that is ubiquitous across the seafloor of the Arctic Ocean, only very steep slopes (>35°) were chosen as dredge sites (Fig. 2). Hundreds of kilograms of angular broken rock fragments collected from the central Chukchi Borderland were plucked from outcrop as recorded by high-tension pulls on the dredge line. High tension on the dredge line occurs when the dredge basket becomes stuck on hard bottom seafloor (Fig. 3A).

To check the conclusion that rock fragments in the dredge were broken from the submarine outcrops during high-tension pulls, we carried out a detailed lithologic comparison of the fragments with the gravel through boulder-size IRD that was also collected during normal dredging. We classify IRD after Huggett and Kidd (1983) by their wedge shape, the presence of glacial striations and polish, roundness, and manganese stain development. Rocks we infer to have been broken from outcrop have approximately uniform thicknesses of manganese crust on surfaces exposed to seawater (white arrow, Fig. 3B), while broken surfaces are fresh and free of manganese precipitate. In contrast, manganese staining on IRD is either nonexistent or of heterogeneous thicknesses due to variable time exposed to seawater after deposition by ice rafting (Fig. 3C). No IRD lithology matched any of the angular and broken rocks that were dredged from outcrop (Fig. 3D).

The high-tension dredge pulls achieved in dredge HLY0905-DS5 yielded broken fragments of mafic amphibolite, garnet-bearing feldspathic gneiss, and potassium feldspar (K-feldspar) augen-orthogneiss. Despite the range of lithologies represented within the central Chukchi Borderland dredge haul, all of the rock fragments have been variably mylonitized and share a lower amphibolite-facies overprint. This observation supports our conclusion that the rocks reported on here were indeed collected from submarine outcrops.

Mafic rocks make up about a quarter of the HLY0905-DS5 dredge haul. Amphibolite samples are blocky and dark gray to blackish green in hand specimens with iron-oxide staining on fractured surfaces (Figs. 4A and 4C). Many samples have deformed quartz veins or quartzofeldspathic-rich bands (Fig. 4C). Two representative samples (#5-009; IGSN ECS008102 and #5-043; IGSN ECS008136) are reported on here.

Sample #5-009 (Fig. 4B) is a medium-grained, well-foliated hornblende plagioclase amphibolite. Amphibole-rich layers contain large (500 µm) relict clinopyroxenes (?) that are fully replaced by chlorite and serpentine minerals. One- to 2-mm-thick lenses of highly sericitized plagioclase occur parallel to foliation and contain little amphibole. Brittle-spaced fractures within the rock contain chlorite alteration. Zircon is visible in the plagioclase-rich compositional domains.

Sample #5-043 (Fig. 4C) is a biotite amphibolite that contains a ∼1-cm-thick leucogranitic seam that was extracted from the sample and analyzed separately. The leucogranite (#5-043L) is composed of recrystallized quartz (35%), intensely deformed and highly sericitized plagioclase (35%), and potassium feldspar (25%) that is less altered but significantly deformed. Chloritized biotite (5%) is disseminated throughout. Apatite and zircon are visible as accessory minerals. In thin section, the contact with the leucocratic seam is very sharp (Fig. 4D). A 2–3 mm biotite-rich reaction selvage is present at the interface. The biotite is aligned parallel or subparallel to the leucogranitic seam, and much of it has altered to chlorite. About 0.5 cm away from the seam, amphibole and plagioclase become the dominant phases, although some biotite persists. Plagioclase exhibits lamellar twining, although most is extensively replaced with sericite.

Garnet-bearing gneisses make up about half of the HLY0905-DS5 dredge haul. These rocks are very finely layered with mm-size altered garnet in hand samples (Fig. 4E). Most of the samples are also fractured at high angles to the compositional banding.

Sample #5-019 (IGSN ECS008044) (Fig. 4F) is composed of strongly recrystallized quartz (35%) with grain sizes that range from ∼100 µm down to 10–40 µm. Thin dark bands are composed of biotite variably altered to chlorite. Feldspars include highly sericitized, 200–400 µm plagioclase phenocrysts (40%) and subordinate microcline (10%). There are trace amounts of chloritized garnet (2%), small amounts of zoisite, secondary muscovite, abundant apatite, and some zircon. Spaced fractures are filled with chlorite.

Sample #5-020 (IGSN ECS008045) is less deformed with overall larger grain sizes relative to the sample described above. Quartz (30%) varies from coarse grained down to a grain size between 20 and 40 µm. The rock also includes highly sericitized plagioclase (40%), potassium feldspar (5%), chloritized biotite (15%), muscovite (2%), and zoisite (2%). Large garnets (1–3 mm) are tabular or fragmental and altered to chlorite. Accessory titanite, apatite, and zircon are visible in the thin section.

The protolith of these quartzofeldspathic rocks is uncertain due to their highly deformed and metamorphosed nature. However, the size and abundance of feldspar, as well as aspects of the U-Pb analysis discussed below, suggest that the protolith of the garnet-bearing feldspathic gneiss was an igneous rock of intermediate composition, although an immature sedimentary protolith cannot be ruled out.

We selected two representative coarse-grained orthogneiss samples from the approximately 20 K-feldspar augen-orthogneiss fragments collected from the HLY0905-DS5 dredge haul. Both samples displayed large (>1 cm) K-feldspar augen visible in hand sample (Fig. 4G). Sample #5-001b (IGSN ECS008094) is composed of recrystallized quartz (35%) with subgrain development and grain boundary migration textures with grain sizes between 50 and 200 µm. Variably sericitized plagioclase (30%) makes up most of the augen with lesser amounts of microcline (15%). The largest augens (>1 cm) are potassium feldspars. Variably chloritized biotite (15%) is aligned and defines the foliation. Abundant zircon was visible in this sample along with large euhedral titanite (250 µ) and apatite.

Sample #5-002a (IGSN ECS008095) is similar in composition, but quartz (30%) is more deformed with grain sizes between ∼50 and 100 µm. The proportion of plagioclase (15%) is subordinate to large K-feldspar augen (35%). There is some chloritization of biotite (10%), which is aligned in a preferred orientation that defined the foliation. Trace epidote, allanite, and abundant apatite and zircon are present.

We employed the sensitive high-resolution ion microprobe–reverse geometry (SHRIMP-RG) located at the Stanford University–U.S. Geological Survey (USGS) Micro Analysis Center to determine U-Pb crystallization ages of zircons in the three lithologies from the HLY0905-DS5 dredge haul. Zircons were concentrated by standard heavy-mineral separation processes and were hand selected and mounted on double-sided adhesive tape on glass plates. Mounts were cast in epoxy, lightly sectioned by grinding with 1500 grit carbide, and ultimately polished using 1-micron polycrystalline diamond. Zircon grains were imaged with a cathodoluminescence (CL) detector mounted within a JEOL 5600 scanning electron microscope (SEM) to identify internal compositional zoning to select spots for U-Pb analyses.

We followed ion microprobe analytical routines modified after Williams (1998). Primary beam tuning produced 4–6 nA O₂-ion beam that created a 25–30 μm diameter, 1–2-μm-deep sputter pit depending upon analysis duration. Both trace elements and U-Pb ion intensities were generally measured in the same run (some trace-element analyses were performed separately on the same mounts, and several samples lacked trace-element data). Analysis times included pre-sputtering of the analysis surface to minimize surface Pb contamination.

Data reduction utilized the SQUID program of Ludwig (2005). Concentration data for U, Th, and all of the measured trace elements were standardized against a well-characterized, homogeneous in-house zircon standard (MAD-green zircon). Measured Pb/U ratios from unknown zircons were standardized using equivalent measurements for R33 standard zircon (Black et al., 2004). Primary beam drift corrections were applied. Zircon U-Pb ages younger than 1 Ga were based on ²⁰⁶Pb/²³⁸U ages calculated from ratios corrected for common lead using age corrected ²⁰⁷Pb as a proxy. Ages older than 1 Ga are based on ²⁰⁷Pb/²⁰⁶Pb ages calculated from ratios corrected for common Pb using measured ²⁰⁴Pb. Common Pb isotopic compositions were estimated from Stacey and Kramers (1975). Age calculations and Tera-Wasserburg diagrams were generated with the Isoplot/Ex program of Ludwig (2003). Zircon U-Pb ages and trace-element data appear in Supplemental Tables A.1–A.5 in the Supplemental File¹.

Whole-rock major- and trace-element compositions were measured for both K-feldspar augen-orthogneiss samples (#5-001b and #5-002a) at the Washington State University (WSU) GeoAnalytical Lab. The samples were crushed, and unaltered rock chips were hand selected for analysis. These chips were pulverized to a very fine powder using an agate bowl at the WSU facility. The rock powder was mixed with di-lithium tetraborate flux (2:1 flux:rock) and fused at 1000 °C in a muffle oven. The resulting bead was reground, refused, and polished on diamond laps to provide a smooth flat analysis surface. The beads were analyzed using the ThermoARL Advant’XP+ sequential X-ray fluorescence spectrometer at the WSU GeoAnalytical Lab. Ten major elements and 19 trace elements were determined. A split of the powder was dissolved using high-strength acids (HF, HCl, and H₂NO₃). The solution was analyzed for trace elements and rare-earth elements (REEs) using the high-resolution, single-collector, inductively coupled plasma mass spectrometer (HR-SC-ICP MS), a Finnigan Element 2, capable of analyzing elements in solution at concentrations as low as parts per quadrillion at the WSU GeoAnalytical Lab.

K-feldspar augen-orthogneiss #5-001b and #5-002a were processed at the Stanford University ICP-MS/thermal ionization mass spectrometer (TIMS) facility using the procedures described in Konstantinou et al. (2013a, 2013b). The sample analyses were compared and normalized relative to 43 analyses of Standard Reference Material (SRM) 987 (average ⁸⁷Sr/⁸⁶Sr value of 0.710144 ± 52 [2σ]; accepted value of ⁸⁷Sr/⁸⁶Sr = 0.71025 ± 1 [2σ]); Balcaen et al., 2005). A secondary standard of BHVO-1 was analyzed seven times and yielded an average value of 0.703483 ± 14 (2σ) (accepted value is 0.703475 ± 17 [2σ]; from Weis et al., 2005) after correction using the normalization factor obtained from SRM 987.

For the Nd isotope analyses, sample ratios were multiplied by a normalization factor determined from JNdi corrections to a value of ¹⁴³Nd/¹⁴⁴Nd = 0.512115 ± 10 (2σ; Tanaka et al., 2000); 26 analyses of JNdi resulted in an average ¹⁴³Nd/¹⁴⁴Nd value of 0.512087 ± 22 (2σ). A secondary standard of BHVO-1 was analyzed three times and yielded an average ¹⁴³Nd/¹⁴⁴Nd value of 0.512969 ± 20 (2σ). This compares to the value for BHVO-1 of ¹⁴³Nd/¹⁴⁴Nd = 0.512986 ± 18 (2σ) from Weis et al. (2005). The results of the Sr and Nd isotope analyses, together with the values used to calculate ⁸⁷Sr/⁸⁶Sri and εNd(i) values, are reported in Supplemental Table A.6 (see ^{footnote 1}).

Whole-rock normalized major elements of the studied K-feldspar augen-orthogneiss (#5-001b and #5-002a) revealed intermediate SiO₂ concentrations (65%–67%), enrichment of Na₂O (3.4%–3.8%), and low K₂O/Na₂O (1.0–1.4; Table 1). Calculated CIPW norms classify sample #5-001b (IGSN ECS008094) as a monzogranite (Q₂₁A₂₆P₅₃) and #5-002a (IGSN ECS008095) as a granodiorite (Q₂₄A₃₁P₄₅) (Supplemental Table A.7 [see ^{footnote 1}]).

Whole-rock trace-element analyses of K-feldspar augen-orthogneiss show an overall enrichment in large ion lithophile elements (LILEs), relative to high field strength elements (HFSE); this enrichment is commonly attributed to the high solubility of LILE in the aqueous fluids abundant in subduction zones (e.g., Rogers and Greenberg, 1990). Dehydration of subducting oceanic crust releases fluids with high concentrations of LILE relative to HFSE, enriching subduction-related magmas in LILE (e.g., Zindler and Hart, 1986). The trace-element patterns of the K-feldspar augen-orthogneisses are similar to I-type granites (Fig. 5A; Supplemental Table A.8 [see ^{footnote 1}]). The Chukchi Borderland rocks demonstrate depletions in Ta, Nb, and Ti, relative to trace elements of similar geochemical behavior, which is also a hallmark of subduction zone magmatism.

Based on radiogenic isotopic analyses, the Chukchi Borderland K-feldspar augen-orthogneiss samples have an εNd value of –4.8 and ⁸⁷Sr/⁸⁶Sr ratios of 0.707. These values are typical of granites emplaced in a continental arc setting or I-type granites (Fig. 5B; Zindler and Hart, 1986) and are not as evolved as typical anatectic S-type granites. Continental arc magmas with ⁸⁷Sr/⁸⁶Sr ratios in the range 0.704 to 0.708 are often interpreted to indicate interaction of mantle-derived melts with continental crust above continental margin subduction zones (Table 1; White, 1979; Pearce, 1983).

Zircon U-Pb results were obtained from two representative Chukchi Borderland amphibolite samples (#5-009 and #5-043). Zircons were also analyzed from the leucocratic seam (#5-043L) discussed above. Zircons from both the amphibolites and leucocratic seam were between ∼150 and 250 µm long, generally anhedral, and yellow to brown on ion probe mounts. Many were highly fractured and contained inclusions. The CL images of zircons from both amphibolites (#5-009 and #5-043) revealed weakly planar, patchy or chaotic zoning, or no zoning (Figs. 6A–6C). These characteristics are typical of metamorphic zircon (Corfu et al., 2003; Yuanbao and Yongfei, 2004).

Ion microprobe U-Pb age results from both amphibolite samples yielded a broad, uniform age distribution (Figs. 6D–6F; Supplemental Table A.1 [see ^{footnote 1}]). Using all measured ages, we calculated weighted mean ²⁰⁶Pb/²³⁸U ages of 508 ± 5 Ma for sample #5-009 (n = 27; Figs. 6D–6G) and 486 ± 20 Ma for sample #5-043 (n = 20; Figs. 6E–6H). Uncertainties are quoted at 95% confidence. The leucocratic seam (#5-043L) also contained zircons exhibiting patchy, planar, or chaotic CL zoning (Fig. 6C). Zircon U-Pb results from the leucocratic seam overlapped those from the amphibolite samples and yielded a weighted mean age of 489 ± 15 Ma (n = 30; Figs. 6F–6I). U-Pb weighted mean ages of all three samples overlap at 95% confidence.

Zircons yielded by the two finely banded garnet-bearing feldspathic gneisses ranged from 50 to 250 µm in length and generally exhibit subhedral to prismatic forms with rounded terminations. Some zircon grains were highly rounded. Zircons ranged in color from yellowish to reddish brown. The most rounded zircons tended to be more reddish color, although this was not consistent throughout. Cathodoluminescence imaging revealed some consistent relationships between zircon shape and internal structure. For example, subhedral grains had oscillatory-zoned cores with dark overgrowths, while prismatic crystals had bright cores and patchy zoned overgrowths (Figs. 7A and 7B). Smaller rounded zircons show patchy or chaotic zoning with no visible core/rim relationships (Figs. 7A and 7B). The U-Pb age distributions for both samples are very similar in that they exhibit age clusters at ca. 480–545 Ma (black and dark red ellipses, Fig. 7D) and 560–650 Ma (red and blue ellipses, Fig. 7D) and older ages between 1100 and 1700 Ma (Fig. 7C; Supplemental Table A.3 [see ^{footnote 1}]).

Zircon CL images reveal that most grains extracted from the K-feldspar augen-orthogneiss samples (#5-001b and #5-002a) are euhedral and exhibit oscillatory zonation with a few texturally distinct cores (Figs. 8A and 8B). Analysis of the oscillatory-zoned regions yielded a range of ²⁰⁴Pb-corrected ²⁰⁶Pb/²³⁸U ages for both samples of between 403 and 465 Ma (Supplemental Table A.4 [see ^{footnote 1}]). The broad spread of concordant ages in these samples yields a mean square of weighted deviates (MSWD) that is inconsistent with a homogeneous population. The youngest age determinations (between ca. 415 and 403 Ma) are discordant and/or have large errors (Fig. 8C, dashed ellipses). The youngest zircons yield trace-element patterns that show enrichment of light rare-earth elements (LREE; Supplemental Table A.5 [see ^{footnote 1}]) that are indicative of hydrothermally altered zircon (Fig. 9C; Hoskin and Schaltegger, 2003). Because these young ages represent a postcrystallization process, they were excluded from calculations of mean ages. All highly discordant analyses were also discarded (Supplemental Table A.4 [see ^{footnote 1}]). After filtering these results, we calculated weighted mean ²⁰⁶Pb/²³⁸U ages of 432 ± 3.8 Ma and 430 ± 3.6 Ma for #5-001b and #5-002a, respectively (Fig. 8C). Nine targeted cores yielded U-Pb ages between ca. 860 and 980, 1100 and 1450, and 1650 and 1825 Ma (Fig. 8D).

Zircon trace-element patterns are controlled by the equilibrium phase assemblage during zircon growth. Garnet is absent from the Chukchi Borderland amphibolites, and feldspar is the stable aluminous phase. The zircon trace-element characteristics of the dredged amphibolite samples are consistent with Eu being preferentially partitioned in plagioclase in that they yield negative Eu anomalies (Fig. 9A; Supplemental Table A.2 [see ^{footnote 1}]). Most zircons recovered from the amphibolite samples appear to have crystallized as new metamorphic grains (see Corfu et al., 2003, for criteria). A few zircons exhibit CL zoning patterns, but these are age equivalent within error of the ages measured from light-colored rims. We thus regard the CL patterns as planar zoning established during metamorphic zircon growth (Figs. 6A–6C). Overall, we interpret the amphibolite U-Pb zircon results in terms of metamorphic recrystallization during a protracted period of deep crustal residence between ca. 508 and 486 Ma.

The strong deformation fabric of the garnet-bearing feldspathic gneisses obscures the nature of their protolith. While their distribution of U-Pb zircon ages is consistent with that expected from detrital zircons from an immature sedimentary protolith, closer inspection of the CL images and zircon trace-element data suggests a more complicated history for these highly deformed rocks.

All of the oldest ages (1100–1700 Ma; Fig. 7C) in both samples (#5-020 and #5-019) are from analyses taken from oscillatory-zoned cores in subeuhedral grains with rounded terminations suggesting that these ages are inherited from an older protolith (Figs. 7A and 7B). The discordance of many of the Proterozoic U-Pb results (Fig. 7C) is attributed to metamorphic overprinting at ca. 500 Ma (see below).

In contrast, the ca. 600 Ma ages were obtained from analyses of either prismatic zircon grains or bright cores and suggest a genetic relationship. We interpret results from the prismatic zircon as recording the crystallization age of a plutonic protolith for the garnet-bearing feldspathic gneisses but acknowledge that the gneisses could have been derived from an immature sedimentary protolith. Weighted mean ²⁰⁶Pb/²³⁸U zircon crystallization ages of 610 ± 24 Ma and 580 ± 20 Ma can be calculated for samples #5-020 and #5-019, respectively (red and blue ellipses in callout, Fig. 7D). Trace-element characteristics of zircon cores and from the prismatic zircons within both garnet-bearing feldspathic gneiss samples display a significant Eu anomaly and HREE enrichment typical of igneous zircon (Fig. 9B; Rubatto, 2002; Hoskin and Schaltegger, 2003; Yuanbao and Yongfei, 2004) (Supplemental Table A.2 [see ^{footnote 1}]).

Finally, all of the 480–545 Ma zircon ages are from either rounded patchy zoned grains or from obvious zircon overgrowth rims that we interpret as metamorphic zircon (gray region, Fig. 7D). These analyses show depletion of HREE, which is consistent with an interpretation of zircon crystallized in equilibrium with garnet (Fig. 9B; Supplemental Table A.2 [see ^{footnote 1}]). The trace-element pattern and ages of the metamorphic zircon overgrowths suggest that these rocks were probably later subjected to the same Cambro-Ordovician metamorphic event that affected the amphibolites collected in the same dredge discussed above.

Zircons from the K-feldspar augen-orthogneiss samples yielded concordant U-Pb ages between ca. 420 and 465 Ma that may represent inheritance from earlier plutonic rocks within a long-lived and deep magmatic system (Fig. 8C). We interpret the mean ages of 432 ± 3.8 Ma and 430 ± 3.6 Ma as approximating the crystallization ages of the K-feldspar augen-orthogneiss samples. Zircons from the augen-orthogneiss samples contain cores that yield ages between ca. 860 and 980, 1100 and 1450, and 1650 and 1825 Ma that we interpret as inherited zircon from a sedimentary protolith (Fig. 8D). These inherited zircon cores have similar ages to the metasedimentary basement in Pearya and western Svalbard (see sections on Pearya Terrane and Western Terrane of Svalbard).

Basement samples dredged from the central Chukchi Borderland (dredge HLY0905-DS5) preserve a record of late Neoproterozoic–Silurian metamorphism and igneous activity that conflicts with previously advanced models that the Chukchi Borderland shared the tectonically quiescent late Neoproterozoic and early Paleozoic passive margin evolution of Arctic Canada (Grantz et al., 1998). Grantz et al.’s (1998) hypothesis was based upon sediment piston-core and box-core samples that were collected along the Northwind Ridge (red X, Fig. 1) of the Chukchi Borderland. There are several possible explanations to account for the discrepancy in geologic history inferred from samples recovered from the two separate locations of the Chukchi Borderland: (1) The unmetamorphosed sedimentary rocks described by Grantz et al. (1998) were deposited atop the basement rocks discovered in the present study, but faulting has caused exposure of different crustal levels across the Chukchi Borderland; (2) an undetected large displacement fault between the two collection sites has juxtaposed genetically unrelated materials; or (3) the Grantz et al. (1998) samples represent distally derived, ice-rafted debris that has no relationship to the Chukchi Borderland.

While it is not at present possible to conclusively select between these possibilities, there are some important considerations. First, the average slope at the collection sites of the Grantz et al. (1998) piston-core study (1997–3720 m water depth) was only 7°–9°. Such gentle slopes are problematic in that they primarily yield IRD when dredged (Figs. 3B and 3C). Second, the samples described by Grantz et al. (1998) were unmetamorphosed. Although this paper only reports on a single dredge haul from the central Chukchi Borderland, four more ECS dredges were collected from the Chukchi Borderland, all of which yielded metamorphic rock from submarine outcrops (Mayer and Armstrong, 2008, 2012). For instance, one ESC dredge haul (cruise number HLY1202; Mayer and Armstrong, 2012) encountered bedrock exposures along a >40° cliff face at the base of the Northwind Ridge less than 35 km north of the sites reported on by Grantz et al. (1998). This HLY1202 dredge yielded ∼300 kg of deformed and metamorphosed manganese-encrusted calcareous sandstones and phyllites (O’Brien et al., 2013). Third, the piston-core samples described by Grantz et al. (1998) contained many of the same lithologies that are present within IRD collected during our dredging operations (Fig. 3D). Finally, features such as manganese crusts were seldom described by Grantz et al. (1998) for the clasts that they studied. Based upon these observations, we speculate that the piston-core samples described by Grantz et al. (1998) consisted primarily of IRD. Consequently, correlation of the Chukchi Borderland to the unmetamorphosed Franklinian passive margin is unsupported by the piston-core sampling.

Below we discuss how the intrusive and metamorphic history of the central Chukchi Borderland basement bears a strong similarity to the history recorded by basement rocks present in the Pearya terrane of northern Ellesmere Island and the western terranes of Svalbard (Brumley et al., 2013) (Fig. 1). This is supported by a growing body of evidence suggesting that these terranes were all likely subjected to late Mesoproterozoic to earliest Neoproterozoic (950–1000 Ma) deformation and anatexis possibly associated with the Grenville orogeny (e.g., Johansson et al., 2005) followed by Neoproterozoic (660–600 Ma) age tectonothermal activity (e.g., Peucat et al., 1989; Majka et al., 2008, 2012) prior to Cambro-Ordovician (540–480 Ma) aged metamorphism and magmatism associated with the amalgamation of these terranes (e.g., Trettin, 1987; Gromet and Gee, 1998). The evidence of Ordovician–Silurian subduction and arc magmatism in all three regions further enhances this correlation.

Both Pearya and the western terranes of Svalbard are thought to record a northern continuation of the Grenville orogen from Greenland into the Arctic region (Trettin, 1991; Johansson et al., 2005; Higgins and Leslie, 2008; Kalsbeek et al., 2008). Metasedimentary rocks of this age in East Greenland were deposited between ca. 1100 and 950 Ma and underwent high-grade metamorphism between ca. 950 and 890 Ma (Strachan et al., 1995). While the protolith ages of the metamorphic rocks are not well known in western Svalbard, they probably mainly consist of Neoproterozoic metasedimentary rocks and intercalated basalts (Gasser and Andresen, 2013). Deformation, metamorphism, and crustal anatexis of these Proterozoic sediments occurred between 1160 and 930 Ma (Peucat et al., 1989; Johansson et al., 2005). In the Pearya terrane, Proterozoic sedimentary and volcanic rocks were deformed, metamorphosed, and intruded by granitic plutons between ca. 1100 and 956 Ma (Trettin and Parrish, 1987). In the Chukchi Borderland dredge samples, zircon cores present within the garnet-bearing feldspathic gneisses are interpreted to be inherited from an older sedimentary and/or metasedimentary protolith and yield U-Pb ages of 1100–1700 Ma. Inherited zircon cores measured in the K-feldspar augen-orthogneisses are 880–980 and 1100–1825 Ma. The inherited zircons found in both of these dredged rock types are consistent with derivation from basement rocks similar to those of Pearya and western Svalbard (Fig. 10).

The significance of the ca. 600 Ma U-Pb ages from the garnet-bearing feldspathic gneisses of the Chukchi Borderland are more difficult to interpret but do suggest another tie to western Svalbard, where Neoproterozoic tectonothermal activity has also been reported (Fig. 7; Peucat et al., 1989; Manecki et al., 1998; Majka et al., 2008, 2012). Zircons from felsic dikes in southwestern Svalbard have revealed U-Pb ages of ca. 670–620 Ma that are speculated to be rift related (Peucat et al., 1989; Gromet and Gee, 1998). Pegmatites of western Svalbard (Majka et al., 2012) are also within error of zircon ages from the garnet-bearing feldspathic gneisses of the Chukchi Borderland. ⁴⁰Ar/³⁹Ar step heating of mineral separates yielded a 616 Ma age for hornblende from the Eimfjellt Group and 585–575 Ma ages for biotite and muscovite from the Isbjørnhamna Group, both from western Svalbard (Manecki et al., 1998). An age of 643 ± 9 Ma was obtained for metamorphic monazite (Majka et al., 2008) providing direct evidence of Neoproterozoic metamorphism in western Svalbard. Majka et al. (2012) also studied zircons and monazite from a pegmatite in southwestern Svalbard, which were dated by U-Pb methods and found to be 651 ± 88 Ma and 675 ± 25 Ma, respectively. Ediacaran ages deduced from U-Pb data on zircons from western Svalbard high-grade rocks (Peucat et al., 1989) have been interpreted as a magmatic age and Cambrian ages (560–500 Ma) as the age of recrystallization near the eclogite facies (Dallmeyer et al., 1989). All of the Neoproterozoic ages from western Svalbard overlap in age with zircon crystallization ages in the dredged garnet-bearing feldspathic gneiss samples from the Chukchi Borderland (600–660 Ma), many of which also displayed Cambrian metamorphic overgrowths. This suggests that the Chukchi Borderland had a Neoproterozoic–Cambrian history similar to that of western Svalbard (Fig. 10). Late Neoproterozoic detrital zircons (630–650 Ma) are also present in Ordovician volcanic sandstone deposits within Pearya (Malone, 2012; Hadlari et al., 2013), which could indicate a connection to source rocks of that age in Svalbard and/or the Chukchi Borderland.

Metamorphic zircon crystallization that occurred between ca. 520 and 470 Ma in the dredged amphibolites and garnet-bearing gneisses of the Chukchi Borderland (Figs. 6 and 7) overlaps in age with tectonothermal events on both Pearya and Svalbard (Fig. 10; Trettin, 1987). On Pearya, Early Ordovician ultramafic rocks, arc-type metavolcanic rocks, and associated metasedimentary rocks of the M’Clintock orogeny are interpreted to represent amalgamation of an arc terrane that was accompanied by metamorphism between ca. 485 and 450 Ma (Trettin, 1987; Trettin et al., 1992; McClelland et al., 2012). In western Svalbard, ⁴⁰Ar/³⁹Ar cooling ages dating micas and whole-rock fractions from blueschists give a Middle Ordovician age (461–475 Ma) for the peak metamorphism, with exhumation dated by an unconformity of Ordovician age (Dallmeyer et al., 1989; Ohta, 1994). This unconformity defines the Eidembreen tectonothermal event of Svalbard (Harland, 1997), which is coeval with the M’Clintock orogeny of Pearya. These ages overlap with analyses of metamorphic zircon from the dredged amphibolites and metamorphic zircon overgrowths from the garnet-bearing gneiss samples from the Chukchi Borderland (Fig. 10), and suggest a possible relationship to the M’Clintock and Eidembreen tectonothermal events.

The dredged K-feldspar augen-orthogneisses from the Chukchi Borderland provide further links to Pearya. Isotopically, the Chukchi Borderland orthogneisses are I-type granitoids and have much lower ⁸⁷Sr/⁸⁶Sr ratios than S-type granitoids (0.720–0.740) typical of collisional settings (Fig. 5B; Pearce, 1983; Rogers and Greenberg, 1990; Chappell and White, 2001; Johansson et al., 2005). The existence of I-type plutons in the basement of the Chukchi Borderland demonstrates that arc magmatism was active in this region during Silurian time. The broad U-Pb age range of concordant zircon exhibited by the orthogneiss samples can be explained by inheritance from wall rocks of a long-lived (Late Ordovician–Silurian) magmatic system (ca. 430–465 Ma). I-type granitic plutons that overlap this age range are present on Pearya (ca. 462–481 Ma; Trettin, 1987). Detrital zircon studies indicate that a magmatic arc continued to be a sediment source in Pearya during Silurian time (Hadlari et al., 2013). Thus the Chukchi Borderland and Pearya both record evidence of Ordovician and Silurian calc-alkaline magmatism. Farther south, younger (ca. 430–410 Ma) S-type granites intruded the Caledonian orogen within East Greenland, Norway, and eastern Svalbard (white dots, Fig. 11). Ordovician-age I-type granitoids (ca. 466 Ma yellow dots, Fig. 11) reappear below latitude 72°N in southeastern Greenland (Kalsbeek et al., 2001; Higgins and Leslie, 2008). We correlate the Late Ordovician I-type granitoids of southeastern Greenland with the Late Ordovician–Silurian I-types of Pearya and the Chukchi Borderland and conclude that Late Ordovician–Silurian arc terranes extended north of the main Caledonian orogen prior to and during the collision of Baltica and Greenland (see Gee and Teben’kov, 2004; Labrousse et al., 2008; McClelland et al., 2012).

All basement rocks dredged from the Chukchi Borderland were variably mylonitized and underwent lower amphibolite-facies metamorphism subsequent to the ca. 430 Ma crystallization ages of the orthogneisses. Large offset strike-slip faults are thought to have translated Pearya and southwestern Svalbard away from the Caledonian collision zone to eventually be amalgamated to the northern edge of Laurentia during Silurian time (Trettin, 1987; Mazur et al., 2009; McClelland et al., 2012). Pearya was accreted to the northern margin of Laurentia in Late Silurian to Early Devonian time (Trettin, 1987; Harland. 1997; McClelland et al., 2012). While some models have suggested that Pearya was previously exotic to Laurentia (Churkin and Trexler, 1980; Trettin, 1987), other reconstructions have preferred a peri-cratonic origin for Pearya (Hadlari et al., 2013 and references therein). Following accretion of Pearya and western Svalbard to northern Laurentia, accounts of contractional deformation in the Yukon (e.g., Lane 2007), the Canadian Arctic (e.g., Piepjohn et al., 2008), north Greenland (e.g., Soper and Higgins, 1990), and Svalbard (e.g., Piepjohn, 2000) have been attributed to the Ellesmerian orogeny. The deformation in this region is thought to have been caused by collision with an unknown northern landmass that has since been rifted away due to the opening of the Amerasia Basin (Embry, 1993). The deformation that produced the mylonitic fabrics in the basement rocks of the Chukchi Borderland may have been related to this younger Ellesmerian orogenic event that postdated the arc-related plutonism in both Pearya and the Chukchi Borderland.

1. The first dredged samples from the Chukchi Borderland were collected during the U.S. Extended Continental Shelf project aboard the icebreaker USCGC Healy (cruise number HLY0905) and offer clear evidence that the borderland is underlain by continental crust. This is important for establishing natural prolongation of the landmass for Extended Continental Shelf claims of Arctic states. Rock samples dredged from the central Chukchi Borderland site include biotite amphibolites, garnet-bearing feldspathic gneisses, and K-feldspar augen-orthogneisses.

2. Amphibolites and associated leucocratic lenses contain only metamorphic zircon, based on their rounded morphology and patchy or chaotic zonation seen in CL images. Zircon U-Pb ages of the metamorphic zircon yield mean ²⁰⁶Pb/²³⁸U ages of 508 ± 5 Ma and 486 ± 20 Ma. A leucogranitic seam extracted from the latter amphibolite yielded an indistinguishable mean U-Pb zircon age of 489 ± 15 Ma. This metamorphic event correlates with the coeval M’Clintock orogeny of Pearya and Eidembreen tectonothermal event of western Svalbard and suggests a genetic relationship of the Chukchi Borderland to these terranes.

3. Two garnet-bearing feldspathic gneisses contain zircon populations of 474–545 Ma, ca. 560–650 Ma, and 1100–1750 Ma. Examination of zircon morphology, internal structure, and trace-element abundances, as well as the abundant large feldspars present, suggest that the protolith of the two gneisses analyzed were intermediate plutonic rocks that crystallized at 610 ± 24 Ma and 580 ± 20 Ma, respectively. An immature sedimentary origin for the protolith of the gneisses, however, cannot be ruled out. The inherited U-Pb ages from zircon cores that range from 1100 to 1750 Ma suggest an affinity with the late Mesoproterozoic to early Neoproterozoic basement rocks of the Pearya terrane of northern Ellesmere Island and the western terranes of Svalbard. U-Pb age analyses of metamorphic zircon and overgrowths (ca. 486–525 Ma) indicate that the gneisses were involved in the same Cambro-Ordovician metamorphic event as the one that affected the amphibolites from the same dredge.

4. The two K-feldspar augen-orthogneisses sampled yield weighted mean zircon ²⁰⁶Pb/²³⁸U crystallization ages of 432 ± 3.8 and 430 ± 3.6 Ma. Based on their geochemistry, they are interpreted as subduction-related (I-type) intrusions.

5. Most tectonic reconstructions of the Amerasia Basin return Chukchi Borderland to a pre-rift position proximal to the northwestern edge of the Canadian Arctic Islands (see Lawver and Scotese, 1990; Embry, 1998; Grantz et al., 1998). We alternatively propose that the Chukchi Borderland was part of an active peri-Laurentian arc terrane(s) that included Pearya and southwest Svalbard that lay outboard of the Franklinian passive margin. If correct, Cretaceous tectonic reconstruction models of the Amerasia Basin should position the pre-rift location of Chukchi Borderland much closer to the Lomonosov Ridge and the Pearya terrane of northern Ellesmere Island (Fig. 11).

The authors would like to thank the United States Geological Survey, Stanford University, and BP for their generous support of this project. The work was also partly supported by NSF Grant No. EAR 0948673. We would like to say thank you to the crew of the USCGC Healy who provided help and support throughout dredging operations, and to Dale Chayes for his dredging expertise. Special thanks to Max Lloyd for his detailed descriptions of the dredged ice rafted debris. We would also like to thank the anonymous reviewers for their constructive suggestions that greatly improved the manuscript. Portions of this study have been presented at the American Geophysical Union annual meeting on 5–9 December 2011 and 9–13 December 2013.