Localized Backarc Extension in an Overall Compressional Setting During the Assembly of Nuna: Geochemical and Isotopic Evidence From Orosirian (1883–1848 Ma) Mafic Magmatism of the Aillik Group, Labrador, Canada

Bimodal volcanism occurs in a range of extensional environments that are characterized by distinctive but overlapping, lithogeochemical and isotopic signatures and lithological assemblages. Where basaltic magmatism is associated with ferroan (A‐type) rhyolites, it is typically the lithogeochemistry and isotopic systematics of the basaltic rocks that are the most beneficial in constraining the tectonic evolution. This study presents geochemical and isotopic data from mafic volcanic rocks of the bimodal Aillik Group, that formed during the assembly of the Paleoproterozoic supercontinent Nuna. Lithogeochemical signatures define two suites. Suite 1 samples have N‐MORB (Normal‐Mid‐Ocean Ridge Basalt) chemical affinities such as flat HREE, and smooth but flat LREE and have εNd(t = 1860 Ma) values between +2.8 and +4.8. Suite 2 samples typically have compositions between N‐MORB and IAT (Island Arc Tholeiites) basalts, with variable REE patterns ranging from IAT to OIB (Ocean Island Basalt). The basalts display moderately developed subduction zone signatures, such as negative Nb and Ti, and have εNd(t = 1860 Ma) ranging from −3.4 to +2.2. Geochemical variability within the Aillik Group mafic rocks is explained by processes active in arc settings and involves mixing between depleted mantle components with variable contributions from crustal and subducted‐slab sources. Coupled with field evidence, the geochemical and isotopic data support emplacement of the basalt in an extensional, backarc basin setting. This backarc basin formed due to localized extension during Nuna assembly. The Makkovik Province preserves a complex orogen characterized by multiple diachronous, compressional, and extensional events along a long‐lived active margin of the Archean North Atlantic Craton.

Several workers have attributed the formation of bimodal volcanic suites to mantle plume activities (Jordan et al., 2004;Li et al., 2005;Li et al., 2002). In contrast, in a rift environment, the formation of bimodal suites is often attributed to the melting of the mantle, resulting in melts that underplate the lower crust, which, in turn, results in the melting of the crust through the transfer of heat and fluids. In extensional settings, melts derived from the mantle and the overlying crust synchronously erupt creating bimodal suites (Ahmad et al., 2009;Barbarin, 1999;Cox, 1988;Rudnick, 1990;Terentiev et al., 2017). In subduction zone environments, a three-component mixing model comprising the mantle wedge, subducting slab, and overlying sediments has been suggested as the mechanism to form bimodal suites (Ellam & Hawkesworth, 1988;Hawkesworth et al., 1991;Saunders et al., 1991). Lithogeochemistry and isotopic systematics of basaltic rocks can be used to infer the nature of the mantle, the nature of the crust, and if subducted slab components were involved in their genesis.
In modern plate tectonic environments where bimodal volcanism occurs, different tectonic settings are characterized by different lithological assemblages, coupled with geochemical and isotopic signatures and associated ore deposits. However, in the Paleoproterozoic, these features are commonly obscured by later alteration, deformation, and metamorphism, making the determination of the tectonic setting more difficult. This paper presents lithogeochemistry, along with Sm-Nd data for mafic volcanic rocks of the Orosirian (1883-1848 Ma) Aillik Group, Labrador, Canada. These rocks are of interest because they provide examples of the early magmatic products of rifting during the initiation of backarc spreading. In addition, the chemical features of this suite of rocks allow insights into the crustal architecture of the growing backarc basin during the assembly of the supercontinent Nuna.
The supercontinent Nuna (also known as Columbia, see Evans, 2013), largely formed between 2.2 and 1.8 Ga, with most of the continental crust growth occurred in a relatively short period between 1.9 and 1.8 Ga (Condie & Aster, 2013;Pehrsson et al., 2016;Zhao et al., 2002Zhao et al., , 2004. The assembly of Nuna featured long-lived, internal ocean closure (intra-oceanic accretion) as various proto-continents collided, creating localized extension in an overall compressive regime due to the multitude of subducting plates (Pehrsson et al., 2016). Evidence for such a localized extensional event is preserved in the bimodal volcanic rocks of the Makkovik orogeny that developed along the margin of Laurentia from 1.9 to 1.7 Ga (Figure 1). The geochemistry of mafic volcanic rocks from the Aillik Group is used to constrain the tectonic setting of the HINCHEY 10.1029/2020EA001489 2 of 30  Gorbatschev and Bogdanova (1993). early basin, and to better define the pre-deformation geotectonic history of the Paleoproterozoic Makkovik Province and its implication for the local development of Nuna.

Regional Geology
In Labrador, part of the assembly of Nuna is recorded along the southern margin of the Archean North Atlantic Craton (NAC; Figure 2). During the assembly of the Makkovikian orogen, juvenile terranes were accreted to the NAC during a protracted period of collisional events along this active margin; inferred to be contemporaneous with the Ketilidian Orogen of southern Greenland (Figure 2). The southern margin of the NAC was a passive continental margin from ca. 2240 to 2130 Ma , but little evidence exists for the drift phase until the area became the locus of subduction, arc magmatism, and juvenile crustal accretion in an overall transpressional continental margin arc setting between ca. 1900 and 1790 Ma (Bagas et al., 2020;Garde et al., 2002;Hinchey et al., 2020;LaFlamme et al., 2013).  The Makkovik Province has been divided into three domains termed, from northwest to southeast, the Kaipokok, Aillik, and Cape Harrison domains (Kerr et al., 1996). However, based on recent bedrock mapping, in combination with geochronological, petrological, lithogeochemical, and isotopic data, Hinchey (2021a) suggested that the boundary between the Aillik and Cape Harrison domains is likely arbitrary and that the domains should be merged. Hinchey (2021a) proposed that the Makkovik Province should maintain the division of the Kaipokok Domain, as it correlates with the Northern Domain of the Ketilidian Orogen (Bagas et al., 2020). The Aillik and Cape Harrison domains are now considered as one domain, termed the Adlavik Domain (Hinchey, 2021a; Figure 3). This is consistent with recent changes to interpretations of the alongstrike Ketilidian Orogen, where Steenfelt et al. (2016) recognized that the original subdivisions did not accurately represent the boundaries between tectonic domains. The orogen is now subdivided into the Northern, Central, and Southern domains (Steenfelt et al., 2016), a terminology adopted by Bagas et al. (2020) because of its simplistic and non-genetic nature.
The Kaipokok domain consists of: (a) a reworked, ca. 3300-2830 Ma (Ermanovics, 1993), Archean gneiss of the North Atlantic Craton (Kerr et al., 1996) (Kerr et al., 1996;. The Kaipokok domain represents a relict passive margin and the foreland zone of the Makkovik Province (Kerr et al., 1996;Ketchum et al., 2002). Several high-strain shear zones mark the boundary between the Kaipokok and Adlavik domains and collectively comprise the Kaipokok Bay shear zone Ketchum et al., 2002).
The Adlavik Domain is correlative with the Central Domain of the Ketilidian Orogen, although it represents a higher crustal level. The Central Domain consists largely of the Julianehåb Igneous Complex (JIC) and based on U-Pb dating, formed in two stages; first from ca. 1868-1836 Ma and the second from ca. 1818-1799 Ma (Bagas et al., 2020;Garde et al., 2002;McCaffrey et al., 2004;Steenfelt et al., 2016). Recent lithogeochemical and geochronological studies in the Ketilidian Orogen have suggested that the previously proposed models of ca. 1848-1800 Ma subduction are over simplified and that this interval includes multiple deformation and magmatic events with a hiatus in magmatism between ca. 1836 and 1818 Ma (Bagas et al., 2020). Based on recent U-Pb zircon dating in both the Ketilidian and Makkovik orogens (Bagas et al., 2020;Hinchey et al., 2020) that extended the timing of felsic magmatism, Hinchey (2021a) suggested that with further U-Pb age dating in both domains there is a high likelihood that there would be an increase in overlapping ages of felsic magmatism. If this postulation is correct, then as documented in the JIC, the Cape Harrison Arc might also exhibit two stages of magmatism; an older 1883-1848 Ma stage and a younger 1815-1799 Ma stage.

Aillik Group
The bimodal, Orosirian Aillik Group of the Adlavik Domain (Hinchey et al., 2020) represents the most voluminous and best-preserved volcano-sedimentary sequence in the region (Figure 3). The group is composed of variably deformed units dominated by felsic volcanic and siliciclastic sedimentary assemblages and lesser mafic volcanic rocks. The stratigraphy of the Aillik Group is complicated by folding and structural overprinting that results in structural repetition of units (Hinchey, 2021a(Hinchey, , 2021bHinchey et al., 2020). The stratigraphic thickness of the Aillik Group is estimated to be between ∼7 and 13 km (Hinchey, 2013;Hinchey & Laflamme, 2016). U-Pb zircon ages for felsic volcanic units of the Aillik Group constrain the onset of felsic volcanism to 1883 ± 7 Ma and continuing episodically until 1848 ± 2.7 Ma, representing a 35 m.y. interval (Hinchey et al., 2020;LaFlamme et al., 2013). Two peaks in volcanic activity are recorded, the first between 1883 and 1870 Ma and the second during the interval 1866 to 1855 Ma (Hinchey et al., 2020). In addition to extrusive volcanism, coeval hypabyssal porphyritic intrusions are also documented (e.g., the ca. 1875-1857 Ma Measles Point Granite Suite) and are considered plutonic equivalents of the Aillik Group rhyolites (Hinchey, 2021a;. The principal structural elements in the Aillik Group are km-large, upright to overturned, gently plunging, open to tight folds with axial-planar fabrics, associated with upper-greenschist to lower-amphibolite metamorphism Hinchey et al., 2020;Hinchey, 2021b). Two southeast-dipping, foliation parallel, crustal-scale shear zones cut the Aillik Group; the Big Island Shear Zone and the Pomiadluk Point Shear Zone (Figure 3). The Big Island Shear Zone separates a region of northwest-verging folds and thrusts to the west, and contemporaneous upright to northeast-verging folds to the east Hinchey, 2013Hinchey, , 2007Hinchey et al., 2020). These shear zones accommodated shortening during the northwestward thrusting of the Aillik Group. Following an interval of amphibolite-facies sinistral transpression, the Big Island Shear Zone was reactivated with a subsequent greenschist-facies sinistral transpression overprint . The region was subsequently refolded by a regional set of open, east-west-trending folds. This regional deformation (D 3+4 ) event is confined only to the Aillik Group and associated synvolcanic plutons, likely reflecting the northwestward transport of the group .
The timing of thrusting, strike-slip deformation, and the development of the pervasive planer fabric in the area are poorly constrained. The onset of regional deformation postdates ca. 1848 Ma, the youngest known age of felsic magmatism in the Aillik Group (Hinchey et al., 2020). Alternatively, it may postdate ca. 1836 Ma, based on the change in tectonic environment observed in the correlative Ketilidian Orogen, marked by a hiatus in magmatism in the JIC (Bagas et al., 2020). Further supporting evidence for this transport is demonstrated by the abundance of syn-deformational mafic dykes within the group that are absent in the Kaipokok Domain (Hinchey, 2021b). Regional deformation was also active at ca. 1801 Ma the age of the foliated Kennedy Mountain Intrusive Suite that intrudes the Aillik Group (Barr et al., 2007).

Lithology and Petrology
Metamorphosed mafic volcanic rocks occur as thin belts throughout the Adlavik Domain and two lithological units are distinguished: (a) a less abundant, weakly to moderately deformed, fine-grained mafic tuff and (b) a more abundant, fine-to medium-grained, weakly to moderately deformed basalt (Figures 4 and 5). Basalt flows comprise both pillowed and massive flows and are not mappable as separate units at the current scale of mapping (1:50,000). Based on lithogeochemistry and Nd isotopic signatures, the basalt and mafic tuff units are divided into two suites; however, current mapping cannot distinguish between the suites as both comprise units of massive and pillow basalt and minor mafic tuff.

Basalt
Basalt occurs as flows up to 60 m thick but typically range from 10 to 20 m ( Figure 4a). Both massive and pillow basalts are fine-to medium-grained, preserve upper chlorite-to lower amphibolite-facies metamorphic assemblages, and contain a variably developed tectonic fabric defined by the alignment of amphibole and biotite (Figures 4b and 4c). Locally, these rocks display widespread epidote alteration, relict vesicular texture, and quartz-epidote veining. Disseminated pyrite occurs throughout the massive and pillowed basalt sequences. Locally, the basalt contains galena-covellite-molybdenite vugs and nodules ( Figure 4d).
In well-exposed areas, thick homogeneous layers of massive basalt have basal chilled margins likely reflecting primary layering. Some horizons contain flattened nodules, 2-10 cm in length, containing aggregates of calcite, feldspar, and epidote representing relict amygdales that may indicate flow tops. Minor quartz-chlorite-epidote veining locally occurs and may mark relict pillow selvages. In areas of low strain, within the pillow basalt unit, pillows are only weakly deformed, with the long axes varying in length from 15 to 50 cm ( Figure 4c). In places, pillows are defined by epidote-altered selvages that weather preferentially. In areas of higher strain, pillows are generally flattened obscuring facing directions. Less abundant, minor mafic tuff occurs as 2-15 m thick, medium to light gray discontinuous horizons within the massive basalt.
Both massive and pillow basalt are characterized by primary plagioclase and clinopyroxene in lower grade samples ( Figure 4e). The clinopyroxene has been partially replaced by green to brown pleochroic amphibole and lesser biotite. Accessory minerals include titanite and magnetite. Moderate sericite alteration affects the plagioclase grains in some samples. In higher-grade basalt samples, primary clinopyroxene is largely replaced by green to brown pleochroic amphibole and magnetite, characteristics of amphibolite-facies metamorphism ( Figure 4f). Amphibole grain rims are typically replaced by chlorite, indicative of a local greenschist facies metamorphic overprint. Evidence for recrystallization includes granoblastic textures in the matrix.

Mafic Tuff
Mafic tuff is fine-grained, green-gray, matrix-supported containing lithic lapilli and angular-shaped broken crystals, and is commonly interlayered with basalt ( Figure 5a). It is composed of primary plagioclase, clinopyroxene, and hornblende ± biotite. Accessory minerals include titanite and magnetite. Clast size variation is relatively chaotic and the deposits are generally moderately to poorly sorted. The deposits are characterized by high ash content with interlayered thin beds that are lapilli-rich. The lapilli clasts are ameboid-shaped, sub-angular to sub-rounded, lithic lapilli that commonly reach up to 2 cm near the base of individual horizons to 0.5 cm toward the top. The matrix varies from fine-to medium-grained mafic-rich ash and is diffusely laminated. The mafic tuff locally contains 1-to 4-mm-long aligned biotite and/or hornblende that define the foliation. Actinolite and chlorite are locally associated with hornblende. Chlorite  porphyroblasts are 1-2 mm in diameter and occur throughout the unit. Thin, 5-m-wide, metabasaltic flows occur within thicker packages of mafic tuff. In higher metamorphic grade samples, amphibole and magnetite have replaced clinopyroxene, indicating amphibolite-facies grade and locally, amphibole has been replaced, in part, by chlorite, indicative of a greenschist-facies overprint (Figures 5c and 5d). In higher strain areas, mafic minerals are aligned and define a strong bedding parallel foliation ( Figure 5d). Greenschist-facies recrystallization is characterized by granoblastic textures in the matrix. The mafic tuff locally displays epidote alteration and is cut by 2-to 5-cm-thick-quartz and/or calcite veins.

Lithogeochemistry and Neodymium Isotopic Chemistry
Major elements and select trace elements were analyzed by ICP-OES (inductively coupled plasma-emission spectrometry) following lithium borate fusion and multi-acid digestion. Other trace elements, including the rare-earth elements (REE's), were analyzed by a combination of ICP-MS (inductively coupled plasma-mass spectrometry) and INAA (Instrumental Neutron Activation Analysis). Details of geochemical methods for all samples are in Table S1. Previously published data are incorporated for comparison and include one sample of mafic tuff and eight samples of basalt from LaFlamme (2011). For this study, six samples of mafic tuff and 29 samples of basalt were analyzed and the data are provided in the Table S1. Table 1 provides a summary of key major-and trace-element-ratios. A subset of 12 samples were analyzed for Nd isotopic composition (Table 2), and analytical methods for the Nd analyses are in listed in Table S1.

Mafic Volcanic Rocks-Suite 2
Suite 2 comprises three samples of mafic tuff and 19 samples of basalt. These are supplemented by one previous analysis of mafic tuff and four of basalt from LaFlamme (2011). The SiO 2 content of samples ranges from 40 to 52 wt%, MgO from 3.0 to 13.40 wt % and all have loss on ignition (LOI) values <3.6 wt%. On the total alkali-silica classification diagram (Le Maitre et al., 2002), the suite plots largely in the basalt to trachy-basalt fields ( Figure 6a); however, there is some scatter to other fields.

Neodymium Isotopic Chemistry
Neodymium isotopic data are presented in Table 2. Analysis for suite 1 includes two mafic tuffs and three samples of basalt, and suite 2 samples include one mafic tuff and eight samples of basalt. In addition, samples that were analyzed as part of this project and published in LaFlamme (2011) (Winchester & Floyd, 1975; modified by Pearce, 1996). (c) Y/Ti versus Zr/Ti plot for the discrimination of tholeiitic through calc-alkalic affinities, based on the premises of Barrett and MacLean (1999) with diagram modified from Lentz (1999 (Goldstein et al., 1984) or +3.9 (DePaolo, 1981). The range in εNd (t = 1860Ma) values for suite 2 indicates that they were derived from a long-term LREE-enriched mantle source or that they have lithospheric/crustal contamination. In contrast, entirely positive εNd (t = 1860Ma) values for suite 1 samples imply their derivation from a long-term LREE-depleted source.

Metamorphism and Alteration
In order to interpret the magmatic history of the Paleoproterozoic mafic rocks, based on their geochemical and isotopic characteristics, it is necessary to evaluate the effects of metamorphism and alteration on their compositions. Although attempts were made to collect the least altered samples, field and petrographic data for the mafic rocks indicate that they have been affected by regional greenschist-to lower amphibolite-facies metamorphism. Metamorphism and hydrothermal alteration can result in mobilization of some elements on the scale of hand samples, particularly for large ion lithophile elements (LILEs) such as Rb, Ba, Pb, and K (e.g., MacLean, 1990;Whitford et al., 1989). However, studies on the effects of alteration on element mobility and geochemical patterns of igneous suites have suggested that select major elements (Al, Ti, Fe, and P), high field strength elements (HFSE; Y, Th, Nb, Ta, Zr, Hf), the rare earth elements (REE; except Ce and Eu) and transition metals (Cr, Ni, Co, Sc, and V) are relatively immobile and typically remain unaffected during alteration and metamorphism (e.g., Jochum et al., 1991;MacLean, 1990;Münker et al., 2004;Sun & Nesbitt, 1978).
To evaluate the degree of alteration, samples are plotted on the Na 2 O versus Al 2 O 3 /Na 2 O alteration index diagram (Figure 8a; Spitz & Darling, 1978). Most samples plot within the fresh to weakly altered field with Na 2 O between 2 and 5% and Al 2 O 3 /Na 2 O of less than 10. Only three samples exhibit albite alteration with Na 2 O values over 5%, and only one sample has less than the threshold of 2% Na 2 O. The samples are plotted in the binary alteration plot (Large et al., 2001) which relates two different alteration indices (AI; Ishikawa et al., 1976) that account for feldspar and glass breakdown to sericite and chlorite against the chlorite-carbonate-pyrite index (CPPI; see Table 1, Figure 8b). Most samples plot in the least altered box for basalts and andesites, with five samples plotting in the dacite field. It is apparent that the rocks are relatively unaltered and preserve primary lithogeochemical signatures; thus, all geochemical data can be used to evaluate the petrogenetic processes and tectonic setting of the mafic volcanic rocks.

Petrogenesis and Magma Source Components
The suite 1 rocks have N-MORB chemical affinities, such as low TiO 2 concentrations, flat HREE profiles, flat LREE and ΔNb ranging from negative to positive values (Figures 6 and 7; Table 1). The overall geochemical signature is consistent with their derivation from a variably depleted/enriched mantle source within a non-arc or backarc setting.  Goldstein et al., 1984]. This supports the proposal that suite 1 rocks are derived from a long term LREE-depleted source (Sm/Nd > CHUR).
Suite 2 mafic rocks have much greater variability in their chemical signatures relative to those of suite one. They typically plot between N-MORB to IAT basalts, with TiO 2 concentrations that range from low to high (0.5-2 wt%), with variably enriched LREE and HREE depletions ranging from IAT to OIB basalts (Figures 6  and 7; Table 1). The suite 2 volcanic rocks have moderately developed subduction zone signatures such as negative Nb and Ti (Figure 7). Their overall geochemical signatures are consistent with derivation from a depleted mantle, with indications of contributions from crustal contamination and/or slab dehydration/ melting processes. Isotopically, the suite 2 samples range from εNd (t = 1860 Ma) = −3.4 to +2.2, ranging to lower values than those for suite 1 rocks, and again indicating a role for lithospheric contamination and/or slab influences in the genesis of this suite. The lack of Nd isotopic data for (a) the (unknown) basement to the Aillik Group or (b) any of the sedimentary rocks of the Aillik, Post Hill, and Moran Lake groups, prevents quantitative modeling of the amount of crustal contamination required to produce the εNd value in the suite 2 basalts.
Many discrimination diagrams for mafic volcanic rocks are inadequate for distinguishing between basalts with MORB or backarc-basin affinities in extensional settings (e.g., Pearce, 1982;Shervais, 1982;Wood, 1980). This is because backarc basins may show arc signatures, such as depletion in Nb-Ta and TiO 2 and they may also reflect MORB-like basalts that have been influenced by lithospheric contamination. Li et al. (2015) noted that many trace-element diagrams fail to discriminate between not only backarc-basin basalt and mid-ocean ridge basalt, but also between continental flood basalt and oceanic plateau basalt, and between different types of arc basalt (intra-oceanic, island, and continental arcs). Only ocean island basalt and some mid-ocean ridge basalt are generally distinguishable in many diagrams (cf. Shervais, 1982;Wood, 1980). Nevertheless, some diagrams can help highlight differences between subduction zone effects and those resulting from crustal contamination. For example, the systematics of the immobile, incompatible trace elements Zr, Nb, and Yb, are largely unaffected by crustal contamination and slab metasomatism, are considered more reflective of mantle sources and are used to elucidate such contributions (Pearce, 1983;Pearce & Peate, 1995).
Yb, and subduction-zone-enrichment signatures to other samples having both subduction-zone and crustal-contamination-signatures (Figure 8c). On the Zr/Y versus Nb/Yb diagram, the mafic volcanic rocks of both suites straddle the ΔNb = 0 line, having values ranging from comparable to depleted mantle (ΔNb<0), to being similar to enriched mantle (ΔNb>0: Figure 8d). This is consistent with the interpretation that the volcanic rocks formed in a backarc basin.
The Aillik Group mafic volcanic rocks variably plot between the N-MORB and IAT end members (Figures 6-8), indicating that the suites are derived from a variable mixture of incompatible-element enriched and depleted mantle. In addition, the suite 2 rocks have the added influence of a subducted slab component likely via hydrous melting, which has increased the La and Th in the mantle source (see Hermann et al., 2006). Although there is a consistent linear array in the Nb/Yb versus Zr/Yb diagram (Figure 8e), both suites contain samples that exhibit negative Nb anomalies (i.e., Nb/Th mn < 1), and also exhibit flat to positive Nb anomalies (i.e., Nb/Th mn ≥ 1; Figure 8f). Negative Nb, Ta, and Ti anomalies have been interpreted as an "arc" signature that reflects inheritance from metasomatism of the overlying mantle wedge by a subducting slab (e.g., Pearce, 1983;Pearce & Peate, 1995  might be interpreted to reflect contamination of ascending mafic magmas by crust, or a combination of both subduction-related mantle-wedge metasomatism and contamination processes. The geochemical variability observed within Aillik Group mafic volcanic rocks can be explained by several processes that are active in an arc setting, although much of this diversity is explained by mixing between depleted mantle source components and variable contributions from crustal contaminants and subduction-modified mantle-wedge sources. The suite 1 basalts have low La/Yb ratios which are interpreted to reflect a melting environment dominated by a high degree of melt and/or reflecting spinel as the primary residual phase (Zhang et al., 2008), whereas the suite 2 samples have La/Yb ratios that vary from low to high making it difficult to ascertain the melting environment. In addition, for the suite 1 basalts the relatively low (La/Sm) and (Tb/Yb), in combination, with the flat MREE and HREE signatures, supports their derivation by a high degree of partial melting in the mantle source. This signature is typical of mantle sources in the spinel facies field and thus indicates depths shallower than 60-70 km (Aldanmaz et al., 2000). crust ( Figure 9); as such the negative Nb anomaly is likely a result of a mantle-wedge subduction input. For most samples, however, increasing La (pm) /Sm (pm) with decreasing εNd (0) values for suite 2 rocks, supports the interpretation that the more negative εNd (t = 1860 Ma) values for suite 2 rocks are a result of lithospheric contamination (Figure 9d). Overall, the data imply that the suite 2 samples were influenced by crustal contamination and to a lesser extent some of the suite 1 samples. In addition, at least a proportion of suite 1 samples have authentic, subducted-slab ("arc") signatures that are not a function of crustal contamination.
Saccani (2015) proposed a new discrimination diagram based on normalized Th and Nb concentrations from ophiolitic rocks (Figure 10). Basalts that form in backarc settings cannot be easily identified using this diagram due to the chemical variability of basalts that formed in this setting. However, once a backarc setting is established using other geological and geochemical constraints, this diagram can help clarify the role of crustal contamination. In Figure 10, the backarc A field represents backarc basin basalt (BABB) characterized by the input of subduction or crustal components (e.g., primitive intra-oceanic or ensialic backarcs), whereas, backarc B field outlines BABB showing no influence of a subduction or crustal component (e.g., mature intra-oceanic backarcs). The suite 1 mafic volcanic samples plot in, or overlap, the boundaries of the backarc B field (Figure 10), whereas the suite 2 samples largely plot in both backarc A and B fields, with some samples showing higher Th values and therefore interpreted to exhibit an increased subduction zone influence ( Figure 10). In addition, the samples follow an assimilation-fractional crystallization trend that is consistent with the isotopic systematics ( Figure 10).
The lithogeochemical and isotopic data support the interpretation that the suite 1 samples represent mostly uncontaminated N-MORB basalts that likely formed later in the development of a continental backarc basin, and that the suite 2 samples represent an earlier pulse that interacted with older crustal/lithospheric components. This chemical transition is typical of modern backarc-basin settings, where underplating of basalt, due to extension, results in high-temperature, crustal anatexis generating felsic volcanism with associated basalt that is variably contaminated. Alternatively, felsic magmatism may result from extensive fractional crystallization from a common mantle-derived parental mafic end-member, coupled with crustal contamination. Early basalt magmatism exhibit IAT signatures, whereas the later stage basalt show closer similarities to MORB (e.g., Gribble et al., 1998;Hickey-Vargas et al., 1995;Pearce et al., 2001;Pearce & Stern, 2006). Alternatively, the initial magmatism in the backarc is dominated by low-degree alkali melts from the lithospheric mantle, which will form OIB-like suites that upon intrusion, interact variably with the continental crust. As extension proceeds in the backarc basin, upwelling of MORB-like backarc asthenosphere results in decompression melting of the mantle to produce basalt with MORB lithogeochemical characteristics (Gribble et al., 1996;Hawkins, 1995).
The continued extension is typically associated with intrusion of mafic magmas at depth, which leads to (a) a weakened lithosphere that enhances extension and basin subsidence and (b) an increased heat flow and/or extensive fractional crystallization resulting in felsic magma generation (see Hinchey, 2021a, and references therein). These early melts are felsic with ferroan (A-type) signatures (high Na 2 O + K 2 O, HFSEs, and REE (except for Eu) values; and their production in significant volumes can act as mid-crustal barriers to the ascent of co-magmatic mantle-derived mafic magmas (Janoušek et al., 2020, and references therein). Ultimately, continued intrusion of mafic magmas enables the felsic barrier to be breached, allowing a more widespread ascent of mafic magma and the onset of basaltic volcanism. The early basaltic magmas tend to be more contaminated than the later pulses because (a) early magmas may well have been stored for significant periods in the crust, increasing the likelihood of crustal contamination, (b) later magmas tend to exploit existing magma chambers/conduits/faults that have been isolated from contamination by earlier crystallization products, and (c) early episodes of basaltic volcanism typically occur through relatively thicker crustal sections.

Field Evidence for an Emerging Backarc Basin
Deposition of the Aillik Group started with a relatively thin package of coarse-grained to conglomeratic, immature clastic sedimentary rocks onto an unidentified basement (Hinchey et al., 2020). The occurrence of immature arkosic sandstone and poorly sorted conglomerate, derived almost entirely from a volcanic source, with interbedded carbonate rocks supports the interpretation that the Aillik Group sedimentary rocks were deposited, at least in part, in a shallow marine environment (Hinchey et al., 2020). The preservation of abundant ripple-marks and cross-bedding supports the interpretation of a shallow-water environment, and the occurrence of mud-cracks indicates local, intermittent, subaerial conditions. The conglomerate unit near Pomiadluk Point (Figure 3), contains foliated granitoid clasts from unknown sources, but the absence of gneissic clasts suggests that NAC Archean basement rocks were not exposed, or were absent when the unit was deposited (Hinchey et al., 2020). Immature sedimentation was followed by felsic volcanism, characterized by atleast 10 km thick pile of alkaline, ferroan (A-type) felsic tuffs, and rhyolites (Hinchey, 2021a) punctuated by the eruption of intermittent basalt magmatism. Sedimentation and deposition of sandstone, conglomerate, and tuffaceous sandstone apparently continued during episodes of volcanic activity. The depositional environment of the Aillik Group, therefore, appears most likely associated with faulting and extension of continental crust, such as in a backarc basin, similar to the modern-day, incipient Okinawa Trough.
The presence of wave-induced sedimentary structures in contemporaneous sedimentary rocks implies volcanism in relatively shallow water (i.e., above wave base) for at least part of the Aillik Group. The pillowed basalts affirm a transition to a subaqueous environment, likely attributed to the drowning of the volcanic apron. The stratigraphy of the Aillik Group is reflective of a rifting backarc environment, where the basin is rapidly subsiding as the arc was being dissected.

Constraints From Felsic Volcanism
In addition to the field evidence for the Aillik Group forming in a backarc basin, the associated felsic volcanic rocks also constrain the tectonic setting. The major-and trace-element lithogeochemical signatures are typical of ferroan (A-type) felsic volcanic rocks and granites, such as high Na 2 O + K 2 O, HFSEs (e.g., Zr, Nb, Ga, and Y), and REE (except for Eu) values, as well as FeO T /(FeO T + MgO) ratios (Hinchey, 2021a). Extended trace-element patterns for these felsic rocks are typical of the compositions of melts derived by partial melting of sialic crust, that have strong negative Nb and Ti anomalies, signatures typically associated with crustal-derived melts related either to subduction or crustal contamination processes (Hinchey, 2021a). However, in A-type end-members of the bimodal magmatic suite, extensive fractional crystallization from a common mantle-derived parental mafic end-member, coupled with crustal contamination, can also produce these chemical signatures (Frost et al., 2016, and references therein). Trace-element ratios, such as high La/ Yb (pm) and Zr/Y, are indicative of a low degree of partial melting at high temperatures (∼850°C-900°C) and relatively low pressure where garnet and amphibole are stable in the source (Hinchey, 2021a). Most of the felsic volcanic rocks and syn-volcanic intrusions have εNd (t = 1860 Ma) values that range from +0.1 to −5.2, with T (DM) model ages that range from 2773 to 2160 Ma, suggesting derivation by partial melting of predominantly Neoarchean to Paleoproterozoic sialic crust (Hinchey, 2021a). The coeval formation of Aillik Group basalts and low-pressure, high-temperature felsic melts can occur by decompression mantle upwelling and continued extension and thinning of the subcontinental lithosphere. Those conditions can provide both basaltic volcanism and high-temperature melting of a shallow crustal source. In summary, the geochemical and isotopic characteristics of the Aillik Group felsic volcanic rocks, and the geological field data, support the model of extension in a backarc setting.

Tectonic Assembly
An early phase of Nuna assembly is documented from 2.2 to 2.0 Ga (Pehrsson et al., 2016); whereas, the southern margin of the NAC was a passive continental margin from ca. 2.24-2.13 Ga . During this interval, the earliest events within the reworked portion of the NAC that forms the Kaipokok Domain are the ca. 2235 Ma rift-related emplacement of the northeast-trending Kikkertavak diabase dyke swarm (Ermanovics, 1993). These dykes are broadly perpendicular to the evolving continental margin. The earliest record of volcanism associated with continental rifting is the ca. 2180 Ma eruption of the protolith to the Post Hill amphibolite, interpreted as transitional oceanic crust based on its age and lithogeochemistry . The nature of the substrate onto which the protolith of the Post Hill amphibolite was deposited is unknown. Ketchum et al. (2002) suggested deposition on a sedimentary substrate such as would have formed in an incompletely rifted continental margin. Alternatively,  suggested an allochthonous substrate requiring no relationship with the rifted margin.
The Makkovik Orogen preserves little evidence of the >165 m.y. drift phase during the initial phases of Nuna assembly (Figure 11a). The pelitic and calcareous rocks in the Drunken Harbor supracrustal belt may have been deposited during this time period . In addition, detrital zircon ages from a Post Hill Group psammite that range in age from ca. 2.15 to 2.0 Ga with subordinate Archean grains (2.7-2.5 Ga), suggest that the Paleoproterozoic zircons may have been derived from an approaching terrane .
Within the Makkovik Province, the earliest documented evidence of supercontinent assembly (Nuna) is compressional tectonism that resulted in D 1 thin-skinned thrusting, development of a regional fabric, and broadly coeval amphibolite-facies metamorphism in the Kaipokok Domain (Figure 11b; Ketchum et al., 2002). This D 1 event is interpreted as representing an initial arc-continent (Nuna) collision (i.e., the Kaipokok Arc). Minimum age constraints for the development of D 1 is 1896 ± 6 Ma based on the timing of metamorphism of the Rhyacian (2235 Ma; Cadman et al., 1993) Kikkertavak diabase dykes (Ketchum et al., 1997). Following the initial accretion of the Kaipokok Arc, subduction continued under the NAC and the continental magmatic arc formed (i.e., the calc-alkaline Island Harbor Bay Plutonic Suite-IHBPS; Figures 11c).  suggested that between 1895 and 1870 Ma, the IHBPS intruded in a dextral transpressive regime associated with regional D 2 deformation. Hinchey (2021a) proposed that D 1 accretion occurred prior to 1896 Ma and that the IHBPS formed following slab-break-off beneath the Kaipokok Domain based on lithogeochemistry of the IHBPS that has depleted HREE and enriched LREE Hinchey, 2021a). As such, the initial arc (i.e., Kaipokok Arc) would have been accreted and slab breakoff would have occurred by ca. 1895 Ma (Figure 11c). Subduction may have continued until 1870 Ma, post-slab breakoff, as there is evidence that the Kanairiktok and Kaipokok Bay shear zones were active as D 2 structures prior to 1870 Ma, and part of IHBPS were emplaced synkinematically into these shear zones (Ketchum et al., 1997). Without the accretion of an early arc, that is, the Kaipokok Arc, the D 1 thin-skinned thrusting, and associated amphibolite-facies metamorphism are difficult to explain. If S 1 is an older fabric (pre-Makkovikian orogeny), the intrusion of the IHBPS would still require a different tectonic setting than that of the Aillik Group; unless the IHBPS actually formed in an extensional regime, however, evidence to support this proposal is lacking (Hinchey, 2021a).
Previous models proposed that the onset of backarc felsic volcanism began by 1883 Ma, prior to the terminal phases of the IHBPS, and suggested that this may reflect a change from arc to backarc magmatism, potentially related to a slowing of the subduction rate, or due to isolated rifting caused by strike-slip or continental escape (Hinchey et al., 2020). This interpretation is inconsistent with the model of formation of the Aillik Group post arc-accretion due to tectonic loading (see Moumblow et al., 2019). A backarc basin would have begun to form prior to ca. 1883 Ma, the age of the oldest felsic magmatism (Hinchey et al., 2020). Based on the data presented herein, lithospheric extension resulted in upwelling of the asthenosphere, decompression melting, and ponding of basalt, thereby resulting in anataxis of the crust and forming the depocenter for the Aillik Group. This is consistent with felsic volcanic lithogeochemical and isotopic data (Hinchey, 2021a), which indicate high temperature felsic magmatism at shallow crustal levels. It is not easy to explain how the Aillik Group could have formed on the Kaipokok Arc and escaped the D 2 compressional deformation that was ongoing until ca. 1870 Ma, and is preserved within the fabric of the IHBPS as it intruded active D 2 shear zones (Hinchey, 2021b). The Aillik Group formed synchronous with the IHBPS but requires an extensional setting, with shallow level, high temperature magmatism, and upwelling of the mantle occurring prior to ca. 1883 Ma and continuing until 1848 Ma. No geological evidence exists that requires the Aillik Group to be deposited proximal to the Kaipokok Arc.
The development of the Aillik Group on an incipient, continental, backarc basin documents local extension during Nuna assembly. Extension, accompanied by the intrusion of mafic magmas at depth, would have led to a weakened continental crust and subsequent basin subsidence, as well as to increased heat flow generating anatexis indicating extension began prior to ca. 1883 Ma. Continued extension and intrusion of mafic magma into the base of the lithosphere culminated in the onset of basaltic volcanism. Suite 2 basalts formed early in the depositional history as they are more contaminated than subsequent pulses as represented by suite 1 basalts.   Hinchey et al. (2020) suggested that the basement to the Aillik Group was a juvenile Neoarchean/Paleoproterozoic continental crust. This is based, in part, on the ages of inherited zircon grains in felsic magmatic samples, largely ranging from 1921-1883 Ma and with lesser Archean zircons (2.7-2.9 Ga), and in the detrital zircon from a sandstone of the Aillik Group, with peaks dominated at 1860 Ma and 2525 Ma. A conglomerate sample from the Ketilidian orogen has similar detrital zircon peaks at ca. 1972, 1989, 2100, and 2200 Ma with no known provenance source in Greenland (Bagas et al., 2020). This supports the hypothesis that the basement to Aillik Group was not the older NAC, but rather Neoarchean/Paleoproterozoic in age and is, at least in part, the crustal root of the Cape Harrison Arc (Figure 11d).
The accretion of the Cape Harrison Arc to the assembled Kaipokok domain marked the change from an extensional to a compressional tectonic regime during Nuna assembly (Figure 11e). This event occurred post-1848 Ma, based on the youngest age of felsic volcanism in the Aillik Group (Hinchey et al., 2020 This accretionary event resulted in the inversion of the Aillik basin as it was transported northwestward on a series of thrust faults, coincident at the surface with the Kaipokok Bay, the Big Island, and the Pomiadluk shear zones Hinchey, 2021a). This regional D 3 event (D 1 for the Aillik Group) was confined to the Aillik Group and associated synvolcanic plutonic rocks, and marks the onset of regional transpression preserved as thrusting, sinistral shearing, and generation of regional-scale folding (Figure 11f). There is a noted lack of magmatic activity until ca. 1815-1799 Ma granitic plutonism in the Makkovik Province.
Most of the ca. 1800 Ma plutons also have ferroan (A-type) chemical characteristics and were emplaced within, or near, active shear zones, in an overall transpressional, to locally transtensional setting marking continued compressional tectonics during Nuna assembly. This overlaps with significant felsic plutonism from ca. 1818 to 1799 Ma in the JIC and NAC of Greenland. Thus, Bagas et al. (2020) interpreted these granites as forming after docking of the JIC to the NAC. These granites formed in a different tectonic setting than the JIC, and their chemical signatures do not constrain the tectonic setting of the JIC arc. This is similar to criticism Hinchey (2021a, 2021b) noted, for using the neodymium isotopic signature of the Aillik Group within the Adlavik domain to define the suture between the NAC and the Cape Harrison Arc (Kerr & Fryer, 1994;Moumblow et al., 2019). The early part of the Cape Harrisson Arc was accreted prior to the 1815-1799 Ma granite intrusions and the Aillik Group is thrust at least 60 km to its current location. Thus, the lithogeochemical and isotopic signature of the later granites cannot reflect the tectonic setting of the initial phase of the arc.
If the interpretation that the 1815-1799 Ma plutonism occurred post-initial accretion of the Cape Harrison Arc (ca. 1836 Ma), then the emplacement of the 1815-1799 Ma granitoids within active shear zones, in an overall transpressional environment, would require reactivation of thrusting (i.e., D 5 ). Thus, the open north-south trending folds of the Aillik Group, which create map-scale fold interference patterns, represent regional D 4 folds that formed prior to the intrusion of the ca. 1800 Ma granitoids, as they are not affected by this event. This D 4 event is likely a result of a change in geodynamic setting, following tectonic loading or reactivation of thrusting during continued subduction prior to 1818 Ma. D 5 would likely reflect renewed magmatism and subduction by ca. 1818 Ma. This renewed subduction is recorded in the emplacement and high-grade metamorphism of the Cape Harrison Metamorphic Suite at ca. 1815 Ma, with associated deformation lasting until 1798 Ma .
In the Makkovik Province, D 6 is recorded as greenschist-facies deformation and dextral strike-slip reactivation of the Kanairiktok and Kaipokok Bay shear zones with the development of minor dextral shear zones of similar orientation throughout the Kaipokok domain . This regional reactivation of shear zones, coupled with hornblende 40 Ar/ 39 Ar cooling ages ranging from 1740-1720 Ma , suggests a significant tectonothermal event possibly related to delamination of the mantle lithosphere . The D 6 structures are thought to be coeval with 1720-1715 Ma plutonic activity throughout the Makkovik province (Figure 11g; . Subsequent, D 7 deformation is documented as localized brittle faulting and as sinistral movement on centimeter-wide structures that typically developed within older shear zones.  suggested that this event was related to either late-orogenic transpression and/or pre-Labradorian regional extension (i.e., assembly of the supercontinent Rodinia).

Conclusions
The assembly of the supercontinent Nuna featured a long-lived internal ocean closure as various proto-continents collided, creating localized extension in an overall compressive plate tectonic regime. Evidence of localized extension is preserved in the rocks of the Aillik Group.
The assembly of Nuna as documented in the Makkovik Province comprises: (1) D 1-Thin-skinned thrusting, development of a regional fabric, and broadly coeval metamorphism resulting from the initial Kaipokok Arc-continent collision (Figure 11b). It is restricted to the Kaipokok Domain and is older than 1896 Ma (2) D 2-Thick-skinned thrusting, shearing, and continued subduction beneath the NAC. Slab break-off is interpreted as predating the intrusion of the 1895-1870 Ma calc-alkaline Island Harbor Bay Plutonic Suite (IHBPS). The IHBPS intruded into the reworked margin of the NAC. Dextral transpression and subduction continued until 1870 Ma (Figure 11c). D 2 is restricted to the Kaipokok Domain (3) Prior to 1883 Ma, an extension initiated in the basement of the Cape Harrison Arc, created an emerging backarc basin with associated bimodal volcanism and deposition of sedimentary units forming the Aillik Group (Figures 11c and 11d). Felsic magmatism occurs between 1883 and 1848 Ma. Mafic rock lithogeochemistry and isotopic data coupled with field observations provide evidence of this backarc-basin development (4) D 3-Thin-and thick-skinned thrusting, sinistral shearing, and development of a regional penetrative fabric and broadly coeval metamorphism of the Aillik Group results from the accretion of the composite Cape Harrison Arc to the Kaipokok Domain. D 3 marks the change from an extensional to a compressional regime during Nuna assembly (Figure 11e). Thrusting of the Aillik Group likely occurred after ca. 1836 Ma (5) D 4-Refolding of the Aillik Group via regional-scale, open, east-west-trending folds that create mapscale fold interference patterns. This is likely related to a change in geodynamic setting, following tectonic loading or reactivation of thrusting, during continued subduction, prior to 1818 Ma (6) D 5-Reactivation of the Kaipokok Bay Shear Zone via dextral shearing is associated with the development of a foliation in the ca. 1818-1799 Ma granitoid intrusions (Figure 11f). Subduction continued beneath the assembled Adlavik and Kaipokok domains, and the Cape Harrison Metamorphic Complex was emplaced and deformed at high temperature at ca. 1815 Ma (7) D 6-Greenschist-facies deformation and dextral strike-slip reactivation of the Kanairiktok and Kaipokok Bay shear zones coupled with widespread 1740-1720 Ma tectonothermal cooling event possibly related to delamination of the mantle lithosphere ( Figure 11g) (8) D 7-Localized brittle faulting and sinistral movement on centimeter-wide structures that typically developed within older shear zones. Related to either late-orogenic transpression and/or pre-Labradorian regional extension (i.e., the assembly of Rodinia) As demonstrated in the Makkovik Province, the assembly of the southeastern Nuna was a complex and diachronous process with multiple compressional and extensional tectonic events along the long-lived active margin of the Archean North Atlantic Craton. Research is just beginning to unravel the complex evolution and tectonic stages of this composite accretionary orogen.

Data Availability Statement
All data are available from the GSNL Geological Atlas located here: https://geoatlas.gov.nl.ca and in Table S1.