2. Geological setting

2.a. Sverdrup Basin

The Sverdrup Basin (Figure 1) is underlain by an up to 10 km-thick sedimentary succession of Devonian clastic wedge strata that were deformed during the Late Devonian-Early Carboniferous Ellesmerian orogeny (Embry, Reference Embry and Trettin1991). The Ellesmerian orogeny was succeeded by initial rifting of the Sverdrup Basin, beginning in the Carboniferous where up to 5 km of Upper Paleozoic strata accumulated in a shelf to deep-basin setting (Balkwill, Reference Balkwill1978), and subsequently underwent significant subsidence in the Triassic followed by Early Jurassic – Early Cretaceous rifting resulting in the opening of the Amerasia Basin (e.g. Grantz et al. Reference Grantz, Eittreim and Dinter1979, Reference Grantz, Hart and Childers2011; Houseknecht & Bird, Reference Houseknecht and Bird2011; Hadlari et al. Reference Hadlari, Midwinter, Galloway, Dewing and Durbano2016). Sediment deposited along the northern margin of the basin has been interpreted to be derived from a poorly constrained landmass that lay to the north of the basin based on sediment provenance direction (Embry, Reference Embry2009). While the basin was long considered tectonically quiescent during the Triassic (e.g. Embry & Beauchamp, Reference Embry, Beauchamp and Miall2008), new data support the idea that the basin had tectonic activity, such as the observation of volcanic ash beds (Midwinter et al. Reference Midwinter, Hadlari, Davis, Dewing and Arnott2016) and the undefined Triassic faults and Jurassic extensional faults at the eastern end of the Sverdrup Basin, which are parallel to the eastern flank of the Amerasia Basin (Lopez-Mir et al. Reference Lopez-Mir, Schneider and Hülse2018).

Figure 1. (a) Map of the Sverdrup Basin showing the location of detrital-zircon samples from the Heiberg Formation/Group. Surface and sea bottom bedrock geology is from Okulitch (Reference Okulitch1991). Detrital-zircon sample locations are from this study, Midwinter et al. (Reference Midwinter, Hadlari, Davis, Dewing and Arnott2016), Hadlari et al. (Reference Hadlari, Dewing, Matthews, Alonso-Torres and Midwinter2018). See Figures 2 and 3 for stratigraphic locations of samples. (b) Map of the Arctic, depicting the basins and regions, mentioned in the text. Outline of AACM (black line) after Drachev (Reference Drachev2011). Rotation of AACM away from the Canadian margin illustrated by black arrows (after). Red dashed line denotes location of Figure 1a. Red circle indicates the approximate location of Mendeleev rise from the Shamshur Seamount sample collected by Tuchkova et al. (Reference Tuchkova, Shokalsky, Petrov, Sokolov, Sergeev and Moiseev2020). AA, Arctic Alaska; AACM, Arctic Alaska – Chukotka Microplate; Ch, Chukotka; M, Mendeleev rise; WI, Wrangel Island.

The Heiberg Formation in the central and eastern part of the basin is subdivided into three members, Romulus, Fosheim and Remus (Figure 2; Embry, Reference Embry1983a), and is approximately 1 km thick along the eastern margin of the basin (e.g. Midwinter et al. Reference Midwinter, Hadlari and Dewing2022). The lowermost strata in the Heiberg Formation, the Romulus Member, records a coarsening-upward succession of mudstone to fine-grained sandstone from a prodelta to delta plain environment; the overlying Fosheim Member is a coal-bearing and sandstone-rich interval deposited in a mixed alluvial-marine environment. The Romulus Member is more arkosic than the Fosheim Member, which is more quartzose. The Remus Member is a sandstone-rich unit representative of shallow marine deposits (Embry, Reference Embry1983a; Midwinter et al. Reference Midwinter, Hadlari and Dewing2022). The Heiberg Formation is stratigraphically equivalent to the Heiberg Group in the western part of the basin (Embry, Reference Embry1983b).

Figure 2. Triassic to Early Cretaceous stratigraphic chart of the Sverdrup Basin (Embry & Beauchamp, Reference Embry, Beauchamp and Miall2008; Hadlari et al. Reference Hadlari, Midwinter, Galloway, Dewing and Durbano2016) with location and reference of detrital-zircon samples with respect to stratigraphy.

The Heiberg Group is comprised of five formations: the sandstone-rich Skybattle Formation (equivalent to the Romulus Member), overlain by marine mudstone of the Grosvenor Island Formation, then coarsening upward into the deltaic Maclean Strait Formation, which is subsequently overlain by marine shale of the Lougheed Island Formation and ultimately capped by the sandstone-dominant King Christian Formation (Figure 2). As identified from palynological studies (Embry & Suneby, Reference Embry and Suneby1994), the Grosvenor Island Formation contains the Triassic-Jurassic boundary. This basin centre mudstone is correlative with thin marine mudstone in the lower portion of the Fosheim Member (Suneby & Hills, Reference Suneby and Hills1988; also see Hadlari et al. Reference Hadlari, Midwinter, Galloway, Dewing and Durbano2016). Additional evidence for this mudstone being correlative to the Grosvenor Island Formation is the presence of ironstone rubble (Midwinter et al. Reference Midwinter, Hadlari and Dewing2022), as ironstone was observed within the Grosvenor Island Formation along the southwestern margin of the Sverdrup Basin, indicative of a starved shelf (Embry & Johannessen, Reference Embry, Johannessen and Vorren1993).

The Heiberg Formation records the transition of the pre-rift stage into the syn-rift stage of the Sverdrup Basin as shallow marine sandstones prograded across the basin and multiple unconformities within the formation indicate a basin-filled state with the rift onset unconformity likely within the Fosheim Member, previously estimated to be ∼200-190 Ma (Hadlari et al. Reference Hadlari, Midwinter, Galloway, Dewing and Durbano2016) or ∼170 Ma (Embry & Beauchamp, Reference Embry, Beauchamp and Miall2008). This was supported by detrital-zircon geochronology suggesting the King Christian Formation (coeval to the upper Fosheim and Remus members) had a change in sediment source relative to the lower Heiberg Group/Formation due to the onset of rifting to the north of the Sverdrup Basin (Midwinter et al. Reference Midwinter, Hadlari, Davis, Dewing and Arnott2016). This period marked the initial extension of the proto-Amerasia Basin, which could have developed as a retroarc system driven by slab rollback, which in turn rifted the Arctic Alaska-Chukotka Microplate (AACM) from northern Laurentia, thus disrupting the pathway for the northern sediment source to the Sverdrup Basin. This key transition occurred within the sedimentary succession of the Heiberg Formation, which makes samples tied to stratigraphic sections measured in the Blue Mountains and at Depot Point (Midwinter et al. Reference Midwinter, Hadlari and Dewing2022) vital to better constrain the timing of the extension.

4. Detrital-zircon geochronology results

Isotopic measurements yielded 812 dates that passed our filtering criteria for use in this study. Results are presented as normalized probability density and cumulative probability plots in Figure 3. Details of the isotopic results and the MDA calculations are presented in the online Supplementary Material at http://journals.cambridge.org/geo.

Figure 3. Measured sections (a) with locations of detrital-zircon data (b) presented as normalized probability density functions. Sedimentological data and exact locations of the Blue Mountains and Depot Point sections can be found in Midwinter et al. 2021. All U-Pb data from this study, except for the Romulus Member sample from Depot Point (Midwinter et al. Reference Midwinter, Hadlari, Davis, Dewing and Arnott2016).

Sample Lower Romulus – Blue Mtns (SHF3) was collected from the lowermost sandstone of the Romulus Member (97 m, Figure 3) in the Blue Mountains, 1–2 m above the contact with the mudstone-rich Barrow Formation. The sandstone bed is fine-grained and trough and planar cross-bedded. The proportions of age groupings from n = 184 measurements were Archean (5%), Paleoproterozoic (18%), Mesoproterozoic (13%), 700–360 Ma (29%), Carboniferous (14%), Permian (9%) and Triassic (6%) with a near-continuous distribution of dates between 610 Ma and 209 Ma. A geologically important age fraction from 350 Ma to 200 Ma consists of 51 of 184 measurements (28%). The calculated MDA incorporates the five youngest grains and yielded a weighted mean age of 215.2 Ma ± 9.4 (2σ).

Sample Lower Fosheim - Depot Pt. (15-DTA-32-580m) was collected from a fine-grained sandstone bed immediately above the contact with the Romulus Member (580 m, Figure 3). The quartzose sandstone bed with erosional base was above the bioturbated sandstone of the Romulus Member and below a thin coal bed (Figure 3; Midwinter et al. Reference Midwinter, Hadlari and Dewing2022). The majority of the n = 172 measurements were Mesoproterozoic (1600-1000 Ma; 41%) and Paleoproterozoic (2500-1600 Ma; 32%). Lesser fractions were Archean (>2500 Ma; 8%) and 700–360 Ma (13%). The vast majority of measurements filtered out by probability of concordance are Proterozoic, with no apparent ²⁰⁶Pb/²³⁸U ages younger than 350 Ma (discordant or concordant, no ages younger than 350 Ma have been filtered out). The youngest date was 353.7 Ma ± 16.5 (2σ); therefore, there are no detrital-zircon grains in this sample dated between 350 Ma and 200 Ma.

Sample Lower Fosheim – Depot Pt. (15-DTA-32-660m) was collected from a medium-grained, quartzose, trough cross-bedded sandstone approximately 80 m above the contact with the Romulus Member (660 m, Figure 3). This sample yielded n = 246 measurements that passed the filtering criteria, and those that were filtered out are mainly Proterozoic with none younger than 350 Ma. The age spectrum is very similar to the basal lower Fosheim Member sample with the majority of the measurements yielding Mesoproterozoic (43%) and Paleoproterozoic (32%) dates. Other age groupings were Archean (8%) and 700–360 Ma (13%). The youngest date was 406.2 Ma ± 9.7 (2σ); therefore, there are no detrital-zircon grains in this sample dated between 350 Ma and 200 Ma.

Sample Upper Fosheim – Blue Mtns (SHF9) was collected from thinly interbedded very fine-grained quartzose sandstone with carbonaceous mudstone in the Blue Mountains (876 m, Figure 3). Additional structures observed in this sedimentary bed were wavy, ripple cross-laminations. This interval is approximately 150 m below the contact to the Remus Member. From n = 210 measurements, 7% were Archean, 36% were Paleoproterozoic, 46% were Mesoproterozoic and 7% were between 700 Ma and 360 Ma. A single grain dated at 261.5 Ma ± 12.2 (2σ) was found between 350 Ma and 200 Ma, and none are Triassic.

5. Discussion

The collection of samples from the Heiberg Formation indicates a vertical stratigraphic transition between two provenance signatures (Figure 4). All detrital-zircon samples of the Romulus Member have a consistent and continuous age spectrum from the Carboniferous to Permian through the Triassic where the youngest-aged fraction is near the age of deposition. For example, the 350 Ma to 200 Ma age fraction in sample SHF3 composes 28% of the detrital zircon (51 of 184 measurements). Conversely, above the contact between the Romulus and Fosheim members, the detrital-zircon signature changes to a spectrum with almost no Carboniferous-Triassic age probability (1 of 628 measurements; no Triassic) that is similar to the Devonian clastic wedge reference spectrum from the Franklinian Basin that has a moderate amount of 700–360 Ma ages, a prominent peak at ca. 430–410 Ma and abundant Meso- and Paleoproterozoic ages (Figure 4). This change in detrital-zircon age signature corresponds to the lithological change from arkosic in the Romulus Member to quartzose sandstone in the Fosheim Member. Data presented here for the Fosheim Member are consistent with U-Pb detrital-zircon data also from the Fosheim Member presented by Pointon et al. (Reference Pointon, Smyth, Omma, Morton, Schneider, Hulse, Rippington, Lopez-Mir, Crowley, Millar, Whitehouse, Frei, Scott and Flowerdew2023). Given that the Fosheim Member along the northern portion of the basin has only the older recycled detrital-zircon age spectrum, it can be interpreted that a shift in provenance likely occurs at the boundary between the Romulus and Fosheim members. Based on palynology from Ellesmere Island and interpreted by correlation of the member boundary to our sections, the Romulus Member is assigned a Norian to late Rhaetian age and the basal Fosheim Member consists of latest Rhaetian strata succeeded by the Triassic-Jurassic boundary interval (Suneby & Hills, Reference Suneby and Hills1988). In the southern portion of the Sverdrup Basin, the correlative Triassic-Jurassic boundary interval lies within the Grosvenor Formation (Embry & Suneby, Reference Embry and Suneby1994), and we speculate from regional correlations that the interval of the lower Fosheim Member from 588 to 607 m in the Depot Point section contains the Triassic-Jurassic boundary (Midwinter et al. Reference Midwinter, Hadlari and Dewing2022). This would place the change in provenance and rift onset unconformity as younger than Rhaetian fossils within the Romulus Member, older than the deposition of the Fosheim Member starting in the latest Rhaetian, and therefore late Rhaetian. Our estimation of the timing is dependent upon stratigraphic correlation from the sections with biostratigraphy to wells logs and our outcrop sections, which is a potential source of error and would benefit from additional geological age data.

Figure 4. Compiled normalized probability (a) and cumulative probability (b) plots with corresponding number of analyses. Data sources: the Devonian clastic wedge reference spectrum (Anfinson et al. Reference Anfinson, Leier, Embry and Dewing2012a, 2012b); Triassic sediment of Svalbard (Bue & Andresen, Reference Bue and Andresen2014); Triassic sediment of Chukotka (Miller et al. Reference Miller, Toro, Gehrels, Amato, Prokopiev, Tuchkova, Akinin, Dumitru, Moore and Cecile2006; Tuchkova et al. Reference Tuchkova, Sokolov, Khudoley, Verzhbitsky, Hayasaka and Moiseev2011; Amato et al. Reference Amato, Toro, Akinin, Hampton, Salnikov and Tuchkova2015); Triassic sediment of Wrangel Island (Miller et al. Reference Miller, Gehrels, Pease and Sokolov2010); Romulus Mbr. (Midwinter et al. Reference Midwinter, Hadlari, Davis, Dewing and Arnott2016; Hadlari et al. Reference Hadlari, Dewing, Matthews, Alonso-Torres and Midwinter2018; this study), Fosheim Mbr./King Christian Fm. (Midwinter et al. Reference Midwinter, Hadlari, Davis, Dewing and Arnott2016; this study). ROU: Rift onset unconformity.

These two distinct assemblages, a recycled source and a volcanic arc source, are exhibited in Triassic strata in sedimentary basins along the western and northern Laurentian margin, including the Sverdrup Basin, Arctic Alaska, Yukon-Tanana, Southern Canada (Quesnel and Western Interior Basin) and the Southwest U.S (Hadlari et al. Reference Hadlari, Midwinter, Poulton and Matthews2017 and references therein). Triassic sedimentary rocks in Chukotka and Wrangel Island, part of the Arctic Alaska – Chukotka microplate, are potentially representative of the pre-rift source area and were most likely adjacent to the northwestern Canadian Arctic during this period (e.g. Gottlieb et al. Reference Gottlieb, Meisling, Miller and Charles2014; Tuchkova et al. Reference Tuchkova, Shokalsky, Petrov, Sokolov, Sergeev and Moiseev2020). The detrital-zircon signature of these predominantly upper Triassic rocks of Chukotka (Miller et al. Reference Miller, Toro, Gehrels, Amato, Prokopiev, Tuchkova, Akinin, Dumitru, Moore and Cecile2006; Tuchkova et al. Reference Tuchkova, Sokolov, Khudoley, Verzhbitsky, Hayasaka and Moiseev2011; Amato et al. Reference Amato, Toro, Akinin, Hampton, Salnikov and Tuchkova2015) and Wrangel Island (Miller et al. Reference Miller, Gehrels, Pease and Sokolov2010) resemble the Romulus Member detrital-zircon age spectrum (Figure 4). A detrital-zircon sample from the Mendeleev Rise is similar to upper Triassic samples from Chukotka and Wrangel Island, but with a greater proportion of near-depositional detrital-zircon ages (Tuchkova et al. Reference Tuchkova, Shokalsky, Petrov, Sokolov, Sergeev and Moiseev2020). The Permo-Triassic detrital-zircon age peaks are also found in upper Triassic strata on the Barents Sea side of the Amerasia Basin (Klausen et al. Reference Klausen, Muller, Slama and Helland-Hansen2017), in Svalbard (Bue & Andresen, Reference Bue and Andresen2014) and Taimyr (Zhang et al. Reference Zhang, Pease, Skogseid and Wohlgemuth-Ueberwasser2016), where they persist through Jurassic strata. The presence of abundant Permo-Triassic detrital zircon in Jurassic strata on Svalbard and near absence in the Sverdrup Basin indicates that the Sverdrup Basin was no longer in sedimentary communication with Svalbard or the Barents shelf in the Early Jurassic (Figure 5). The comparison of detrital-zircon data does not eliminate the transport of sediment from the Sverdrup Basin to the Barents region, but it does preclude the transport of sediment from the Barents region to the Sverdrup Basin. The interpretation of detrital zircon in the Barents shelf region is of provenance from the northern Urals (e.g., Miller et al. Reference Miller, Soloviev, Prokopiev, Toro, Harris, Kuzmichev and Gehrels2013; Klausen et al. Reference Klausen, Muller, Slama and Helland-Hansen2017; Klausen et al. Reference Klausen, Rismyhr, Muller and Olaussen2022), and so if the Sverdrup Basin received such sediment in the Late Triassic, this flux had ceased by the Early Jurassic.

Figure 5. The early Mesozoic position of Svalbard relative to the Sverdrup Basin, based on the palaeogeographic reconstruction of the Amerasia Basin by Grantz (2011). Detrital-zircon U-Pb age spectra from Lower Jurassic strata of the Sverdrup Basin are compared to Svalbard, note the difference in the Permian-Triassic interval (Sverdrup Basin, Midwinter et al. Reference Midwinter, Hadlari, Davis, Dewing and Arnott2016, and this study; Bue & Andresen, Reference Bue and Andresen2014).

It is possible that part of the provenance change could be due to a cessation of magmatism in the source area, in which case, the 350–215 Ma age fraction would still be present. The virtual absence of the 350–215 Ma fraction indicates that there was an extra-basinal change in sediment routing patterns and even precludes recycling of upper Triassic deposits from the Sverdrup, Barents, Taimyr and the Arctic-Alaska microplate.

Previous work discussed how the Heiberg Formation/Group marks a tectonic transition in the Sverdrup Basin in terms of basin-filling patterns, normal faults and detrital-zircon provenance (Hadlari et al. Reference Hadlari, Midwinter, Galloway, Dewing and Durbano2016; Midwinter et al. Reference Midwinter, Hadlari, Davis, Dewing and Arnott2016). From new U-Pb detrital-zircon dating, we show that the transition occurs at the boundary between the Romulus and Fosheim members. We rely on previous biostratigraphy of the Romulus and Fosheim members to indicate that the timing of this transition is during the latest Triassic (latest Rhaetian). The detrital-zircon age signature of the Fosheim Member is distinct from upper Triassic strata of the Arctic Alaska – Chukotka microplate, and from Jurassic strata of Taimyr and Svalbard. We can therefore state that sediment from the extra-basinal source of 350–200 Ma detrital zircon had no longer reached the Sverdrup Basin in the lower Jurassic. Based on the similarity to the Devonian clastic wedge, we interpret that the detrital-zircon provenance of the Fosheim Member was sedimentary recycling from erosion of relatively local sedimentary basins, such as the Franklinian Basin, which underlies the Sverdrup Basin (cf. Hadlari et al. Reference Hadlari, Swindles, Galloway, Bell, Sulphur, Heaman, Beranek and Fallas2015). Furthermore, we infer that the provenance transition was a response to regional extension (see Hadlari et al. Reference Hadlari, Midwinter, Galloway, Dewing and Durbano2016), which disrupted sediment transport pathways by the formation of extensional basins that trapped sediment, acting as a sediment sink (cf. Withjack et al. Reference Withjack, Schlische and Olsen2002; Oliebrook et al. Reference Oliebrook, Barham, Fitzsimons, Timms, Jiang, Evans and Mcdonald2019). Between the newly-developed proto-Amerasia Basin and the Sverdrup Basin was a palaeo-high, the Sverdrup Rim (Meneley et al. Reference Meneley, Henao, Merritt and Yorath1975), which was a horst that likely prevented sediment from entering the northern part of the Sverdrup Basin and likely acted as a source of sediment for both basins (Figure 6). We interpret that by formation of localized sediment sinks and barriers, the tectonic localization of provenance thereby prevented sediment from the extra-basinal igneous source from reaching the Sverdrup Basin.

Figure 6. Tectonic environment of northern Laurentia adapted from palaeogeographical maps of Plafker & Berg (Reference Plafker, Berg, Plafker and Berg1994), Nokleberg et al. (Reference Nokleberg, Parfenov, Monger, Norton, Chanďuk, Stone, Scotese, Scholl and Fujita2000), and Midwinter et al. (Reference Midwinter, Hadlari, Davis, Dewing and Arnott2016).