Detrital zircons of the vast Triassic Snadd and De Geerdalen formations, Barents Shelf, reveal temporal changes in sediment source

the vast

The provenance of the Late Triassic Sverdrup Basin deposits has also been the topic of different interpretations, one favouring a Uralian Orogen source while another suggests derivation from an island-arc terrane that lay to the north of the Sverdrup Basin (Omma et al., 2011;Midwinter et al., 2016, Anfinson et al., 2016;Hadlari et al., 2018).
New research into the tectonic evolution in the Arctic region helps to identify two important events that potentially explain a Triassic zircon source, namely tectonism in the Novaya Zemlya-Kara Sea and Taimyr regions (Fleming et al., 2016;Klausen et al., 2017;Zhang et al., 2018a;2018b) and the Siberian Traps large igneous province (LIP) (Zhang et al., 2016).Geographically controlled, intra-formational variation in detrital-zircon age spectra has also been documented, suggesting that different parts of this extensive, Triassic depositional system may have been fed by different river systems sampling different sources (Fleming et al., 2016).Flemming et al. (2016) suggested that sediments off the northern margin of Norway were derived from a Caledonian source, whereas sediments farther north on the Barents Shelf were derived from either South Uralian (The greater Ural -mountains region) or North Uralian (the present day Pai-Koi, Novaya Zemlya and Taimyr region) sources.
In this contribution, we present detrital-zircon U-Pb and Lu-Hf data from the Middle-to Upper Triassic Snadd-and De Geerdalen formations on the Barents Shelf, generally interpreted to have been derived from Uralian sources, as defined by Flemming et al. (2016).Our new data come from various intraformational stratigraphic positions and locations not previously analysed to help refine the general geological understanding of a Uralian source.This work questions the validity of contemporaneous South-and North Uralian sources.Instead, these new data suggest that different sediment sources may have been dominant at different times, most likely due to tectonic events in the source regions.
We further show that this depositional system was much more widespread than the Barents Sea area, most likely also including the distant Sverdrup Basin, thus adding important new knowledge about what is arguably the largest known delta depositional system in the world (Klausen et al., 2019;Gilmulina et al., 2021).

Ediacaran to Triassic Arctic geological evolution
The Ediacaran to Triassic evolution of the Barents Sea region (including Arctic Russia, Scandinavia and Greenland) is characterised by several key tectonic events influencing the development and provenance signatures of the Barents Shelf (Fig. 1).The Archaean through Proterozoic cratons of Siberia, Baltica and Laurentia likely represented independent continents in the late Neoproterozoic, after break-up of the supercontinent Rodinia (Cocks & Torsvik, 2007, 2011;Nystuen et al., 2008;Willner et al., 2019).Siberia and Baltica may have been located close to each other, at a distance from Laurentia (Priyatkina et al., 2017).This configuration changed in the Phanerozoic, with collision between Laurentia and Baltica in the Siluro-Devonian Caledonian Orogeny.The later Devonian to Permian collision between Baltica, Kazakhstania and Siberia in the Uralian Orogeny represents the final amalgamation of the northern part of Pangea (Puchkov, 2009;Nikishin et al., 2010;Vernikovsky et al., 2020).Extensive volcanism associated with the formation of the 250 Ma Siberian Traps LIP on the Siberian Craton marked the transition into the Mesozoic Era (Reichow et al., 2009;Ivanov et al., 2018).As the figure represents an extended period, the considerable overlap between units is not accounted for in the figure.Late Triassic palaeogeography, extent of depositional and offshore areas modified after Sømme et al. (2018).Outline of the Timanian and Caledonian Orogens after Gee et al. (2006), Willner et al. (2019) and Ershova et al. (2016), outline of the present-day Ural mountains after Puchkov (2009), outline of the Central Asian Orogenic Belt after Xiao et al. (2015) and Ershova et al. (2016), outline of the Siberian Traps large igneous province after Ivanov et al. (2018) and the outline of the Kara Orogen after Vernikovsky et al. (2020).The South Taimyr event is suggested based on Zhang et al. (2018a).Possible location of Carnian volcanic activity based on Letnikova et al. (2014) and references therein.

Neoproterozoic tectonic events
Few large-scale tectonic events impacted Baltica and Greenland in the Neoproterozoic until continental breakup and plate divergence around 615 Ma (Nystuen et al., 2008).
The Timanian Orogen resulted in metamorphism of passive-margin sedimentary rocks and magmatic activity in an accretionary tectonic setting that peaked between 580 and 550 Ma (Olovyanishnikov et al., 2000;Gee & Pease, 2004), but may have lasted from 700 to 515 Ma (Kuznetsov et al., 2007).
Detrital zircon of Timanian age is predominant in Cambro-Silurian successions in the Novaya Zemlya and Severnaya Zemlya archipelagos (Lorenz et al., 2008(Lorenz et al., , 2013)), and these sediments were likely more widely distributed.The Neoproterozoic Taimyr Orogen, located on the Taimyr Peninsula (Fig. 1), represents a Neoproterozoic accretionary orogen north of the Siberian craton (Priyatkina et al., 2017).Magmatic zircon was produced during two accretionary phases, first at 900-800 Ma related to continental-arc magmatism and at 750-600 Ma in a back-arc to continental-arc setting (Priyatkina et al., 2017).The latter event may have been linked to the Timanian Orogen (Priyatkina et al., 2017).
The Caledonian Orogen covered a large area of what is now East Greenland -western Scandinavia, Svalbard and the Barents Sea (Gee et al., 2006;Corfu et al., 2014).Zircons derived from the Caledonian Orogen typically range in age between 500 and 390 Ma and are commonly accompanied by large populations of Mesoproterozoic grains (Bingen & Solli, 2009;Pózer Bue & Andresen, 2014).
Caledonian syn-and post-collisional sedimentary rocks, especially Devonian Old Red Sandstones, are present in large areas of the Arctic (Dineley, 1975;Lorenz et al., 2008;2013), making them likely candidates for later erosion and re-sedimentation.

Triassic regional development
The timing and tectonic significance of Permo-Triassic events in the Taimyr, Severnaya Zemlya, Kara Sea and Novaya Zemlya region (Fig. 1) remain enigmatic, although several recent scientific contributions have shed light on the topic (Zhang et al., 2016;2018a;2018b;Khudoley et al., 2018;Vernikovsky et al., 2020).In North Taimyr, the Siberian craton collided with the Kara Terrane resulting in the Permian to Middle Triassic Kara Orogeny (Vernikovsky et al., 2020).Deformation is documented in South Taimyr in the Late Triassic (225 Ma; Zhang et al., 2016;2018a) or possibly latest Triassic-Early Jurassic (Khudoley et al., 2018) and resulted in substantial amounts of crustal shortening (Zhang et al., 2018a).Deformation and erosion on Novaya Zemlya are also documented in the Late Triassic at around 210 Ma (Zhang et al., 2018b).
At the Permo-Triassic boundary, formation of the Siberian Traps LIP involved extrusion of large volumes of mostly mafic volcanic material in Siberia (Reichow et al., 2009).The active phase of volcanism seems to have been restricted to approximately three million years, around 250 Ma (Reichow et al., 2009).
Younger Triassic igneous rocks have been documented in northern Siberia in recent years (Letnikova et al., 2014;Ivanov et al., 2018;Vernikovsky et al., 2020), of which trachytic volcanic rocks in the Carnian Osipai Formation in the Lena delta area represent the largest volumes (Letnikova et al., 2014 and references therein).Several researchers have suggested that Middle and Late Triassic magmatism represents a late-stage continuation of the Siberian Traps LIP (Letnikova et al., 2014;Zhang et al., 2016;Vernikovsky et al., 2020).

The Triassic Barents Shelf
In the Early Triassic, the Barents Shelf was a 1.3 million km 2 basin in the Boreal Sea, north of Pangea (Pózer Bue & Andresen, 2014;Klausen et al., 2015;Sømme et al., 2018).In the Lower Triassic succession, there were several proximal depositional systems in and around the basin, as documented proximal to the Baltican Shield and Greenland (e.g., Eide et al., 2018).The predominant sediment source seems, however, to be the Uralian Orogen, with a sediment influx prograding from the southeast (all directions refer to present day) (Mørk, 1999;Riis et al., 2008;Glørstad-Clark et al., 2010;Henriksen et al., 2011;Lundschien et al., 2014;Klausen et al., 2015Klausen et al., , 2019;;Sømme et al., 2018;Harstad et al., 2021).Continuous sediment supply from the east-southeast eventually led to a complete infilling of the basin (Lundschien et al., 2014;Klausen et al., 2015Klausen et al., , 2019)).Deposition of the marginal-marine to fluvial Snadd and De Geerdalen formations -the topic of this contribution -during the Ladinian to lower Norian is considered the final stage in this evolution (Klausen et al., 2015(Klausen et al., , 2019)).The northerly derived Pat Bay Formation in the Sverdrup Basin potentially represents the farthest known extent of this system (Omma et al., 2011), although this interpretation remains disputed (Midwinter et al., 2016;Gilmullina et al., 2022).In the latest Triassic, parts of the region developed an erosional relief (Klausen et al., 2017).
Seismicity-based research interprets the Snadd and De Geerdalen-formation sedimentary rocks to be deposited in mappable, large-scale clinoforms prograding from the southeast in a northwest direction (Riis et al,. 2008;Glørstad-Clark et al., 2010;Lundschien et al., 2014;Klausen et al., 2015), supported by palaeocurrent measurements from Svalbard (Høy & Lundschien, 2011).The most likely source of these sediments is the Uralian Orogen (Riis et al,. 2008;Glørstad-Clark et al., 2010;Lundschien et al., 2014;Klausen et al., 2015), inferred from the transport direction and original proximity to the Uralian mountain belt.A second, much less voluminous Greenland sediment source is reported on the western margin of the basin, supported by palaeocurrent observations on western Svalbard (Mørk et al., 1982) and seismic sections along the western shelf margin (Glørstad-Clark et al., 2010).
Published detrital-zircon age distributions from Svalbard (Pózer Bue & Andresen, 2014), Franz Josef Land (Soloviev et al., 2015;Khudoley et al., 2019) and the southern Barents Sea (Fleming et al., 2016;Flowerdew et al., 2019;Line et al., 2020) show zircon populations mostly in the age range 600-225 Ma.This age range is distinctly different from zircon populations in sediments derived from Greenland and Baltica (Pózer Bue & Andresen, 2014).Key zircon age peaks are observed from various locations and stratigraphic levels within the Snadd and De Geerdalen formations and include a dominant peak at 300 Ma, several small peaks in the range 400-300 Ma, and a minor peak at 550 Ma that is characteristic for samples from the southern Barents Sea (Fleming et al., 2016).Published zircon age spectra from locations on the northern Barents Shelf show a somewhat different pattern with dominant age peaks at 235, 300 and 420 Ma (Pózer Bue & Andresen, 2014;Soloviev et al., 2015).Detrital zircon age spectra variations within the Snadd and De Geerdalen formations have been attributed to geographical variations in provenance (Fleming et al., 2016), including a North Uralian provenance on Svalbard and Franz Josef Land and a South Uralian provenance on the southern Barents Shelf (Fleming et al., 2016).(2021) investigated detrital Cr-spinel compositions on the same samples as those discussed in this contribution and concluded that a similar, uniform source of detrital Cr-spinel was likely for all samples.
They advocated a metamorphosed ophiolitic ultramafic complex as the ultimate source for the Cr-spinel in the investigated samples and tied this source to the Uralian Orogen.This source is distinctly different from contemporaneous Siberian Traps LIP-related detrital Cr-spinel in Siberia (Nikolenko et al., 2018).

Sampling
This study is based on detrital-zircon U-Pb and Lu-Hf data from 11 samples in a SE-NW transect across the Norwegian Barents Shelf (Fig. 2, Table 1).Sample details are presented in Table 1; see Harstad et al. (2021) for further sample details and a discussion of the Cr-spinel data.All samples were collected from sandstones within the Snadd Formation on the Barents Sea Shelf and the equivalent De Geerdalen Formation onshore Svalbard and represent parts of a prograding delta succession deposited in barrier, shoreface, marginal-marine and delta-plain environments (Bugge et al., 2002;Riis et al., 2008;Stensland et al., 2013;Lundschien et al., 2014;Vigran et al., 2014).Different clinoforms (e.g., Klausen et al., 2015)  Basin, Bjarmeland Platform and Sentralbanken High in the Barents Sea, and Hopen, Edgeøya and West Spitsbergen (van Koelenfjorden) onshore Svalbard.The offshore Snadd Formation samples are from shallow stratigraphic cores drilled by IKU (SINTEF Petroleum) and the Norwegian Petroleum Directorate (NPD) (Bugge et al., 2002;Lundschien et al., 2014).Onshore samples of De Geerdalen Formation sandstones were collected during a SINTEF-and NPD-led expedition in August 2014   Bugge et al. (2002), and Vigran et al. (2014), while Lundschien et al. (2014), and Klausen et al. (2015) add lithostratigraphic correlations to the interpretations.

Methods
The collected samples were crushed and dry-sieved to obtain the 37-250 µm fraction.Zircon was separated from the sand fraction by heavy-liquid (diiodmethane, 3.325 g/mL) separation and handpicked with tweezers under a binocular microscope.Grains were mounted in epoxy and polished to expose their interiors.A 1450VP Scanning Electron Microscope (SEM) from LEO Electron Microscopy LTD at the Geological Survey of Norway (NGU), Trondheim, was used to obtain back-scatter electron and cathodoluminescence (CL) images.
At NGU, an Element XR single-collector, high-resolution ICP-MS coupled to a New Wave Research UP193-FX nm short-pulse excimer laser-ablation system was used to measure U-Pb isotopes in zircon.The laser ablated lines up to 60 μm with a spot size of 15 μm, a speed of 2 μm/s, a repetition rate of 10 Hz and an energy fluence of 3-4 J/cm 2 .Background measurement for 29 s was followed by 30 s of ablation with measurements of the following isotopes: 202 Hg, 204 Pb, 206 Pb, 207 Pb, 208 Pb, 232 Th and 238 U. Analyses with high common lead were excluded from further calculations based on the 204 Pb measurements.The natural zircon reference material GJ-1 (608.5 ± 1.5 Ma; Jackson et al., 2004) was applied as a primary standard.Natural zircon material Plešovice (Sláma et al., 2008), 91500 (Wiedenbeck et al., 1995), OS-99-14 (Skår, 2002), Temora (Black et al., 2003) and Z-6412 (unpublished) were routinely analysed to assess data reliability.U-Th-Pb isotope ratios were calculated using Glitter (Van Actherbergh, 2001).
At Curtin University, zircon crystals were analysed using a split-stream laser ablation system where a portion of the ablated material was divided between a quadrupole ICP-MS for U-Pb analysis and a multi-collector ICPMS for Lu-Hf analysis.Zircon was ablated using a Resonetics RESOlution M-50A-LR system, incorporating a COMPex 102 193 nm excimer UV laser.Following two cleaning pulses and a 40 s period of background analysis, samples were spot ablated for 35 s at a 10 Hz repetition rate using a 50 μm beam and laser energy at the sample surface of 2.2 J/cm 2 .An additional 40 s of baseline was collected after ablation.The sample cell was flushed with ultrahigh purity He (300 mL/min) and N 2 (1.0 mL/min) and high purity Ar was employed as the plasma carrier gas.
Zircon crystals from the Mud Tank carbonatite locality were analysed together with the samples in each session to monitor the accuracy of the results.
The 207 Pb/ 206 Pb age is used for zircons older than 1000 Ma, while the 206 Pb/ 238 U age is used for younger zircon grains.The U-Pb and Lu-Hf data are presented using the detzrcr R application (Andersen et al., 2018).All zircon data are provided in Electronic Supplements 1 and 2.

Results
A total of 1428 U-Pb isotope analyses were performed at NGU.

Detrital zircon U-Pb age data
The data are presented according to geographic/geological location and stratigraphic position within the Snadd and De Geerdalen formations (Fig. 3); sample numbers are in parentheses.All analysed samples contain detrital zircons with U-Pb age ranges overlapping with the age of deposition.
Twenty-eight percent of the zircon grains in the dataset range between 2900 and 635 Ma.The age peaks in this older subset are much smaller than for the younger zircon populations and similar between the different samples.In Figure 4 Nordkapp Basin samples (1, 2 and 3), from the southern part of the study area, represent both the lowest and the uppermost stratigraphic levels included in this study (Fig. 3, Table 1).The stratigraphically lowest, Ladinian samples (1 and 2) are dominated by Carboniferous (300 Ma) and Ediacaran -Cambrian peaks (540 Ma) (Fig. 3J, K).The uppermost Carnian Nordkapp Basin sample (3) is distinctly different, with a dominant 235 Ma age peak and a scattered distribution containing Ordovician to Triassic-aged zircons with minor Silurian and Carboniferous peaks (Fig. 3A).
Bjarmeland Platform and Svalbard samples (4, 9 and 10) of middle and upper Carnian age have similar Phanerozoic zircon age distributions with a dominant peak in the Triassic (235 Ma), a Devonian to Permian peak (300 Ma) and an Ordovician to Silurian peak (425 Ma) (Fig. 3C-E).West Spitsbergen sample (11) of upper Carnian/lowest Norian age is distinctly different from other samples on Svalbard, with a much larger Precambrian zircon age population dominated by Mesoproterozoic and early Neoproterozoic ages (Fig. 3B).
There are four notable Triassic to Ediacaran age groups in the dataset (Fig. 3, right-hand column). A

Detrital zircon Lu-Hf isotopic compositions
The investigated detrital-zircon populations display dominantly juvenile (i.e., εHf t >0, superchondritic) Lu-Hf isotopic compositions (Fig. 5), but with a significant number of subchondritic values.There is no systematic difference between the samples.Archaean and Proterozoic zircon grains have varied but  (Ershova et al., 2016;Priyatkina et al., 2017).However, care must be taken when attempting to identify particular continents based solely on detrital-zircon data (Andersen et al., 2016;Slagstad & Kirkland, 2017) mostly positive εHf t values ranging from -10 to +13.Devonian and Carboniferous zircons have εHf t values mainly between CHUR (chondrite-uniform reservoir) and depleted mantle (DM), indicating juvenile crustal addition or reworking of juvenile crust in the source region(s) in this period.The small number of Hf analyses with corresponding U-Pb discordance <10% for grains younger than c. 300 Ma makes it difficult to say much about the evolution of the youngest source regions.Adopting a more conservative approach in which we define the discordance limit to the closest approach of the 2σ error ellipse to the concordia curve, a larger number of analyses are included.In this case, the samples show a distinct trend towards more negative εHf values, indicative of reworking of older, more isotopically evolved crust.

Provenance of the Snadd and De Geerdalen formations
The results from our study are broadly in accordance with previous provenance studies of the Snadd and De Geerdalen formations, suggesting a dominant sediment source located southeast of the basin (Mørk, 1999;Riis et al., 2008;Omma et al., 2011;Pózer Bue & Andresen, 2014;Lundschien et al., 2014;Klausen et al., 2015;Soloviev et al., 2015;Fleming et al., 2016;Flowerdew et al., 2019;Khudoley et al., 2019;Harstad et al., 2021).However, the addition of new data from lower Carnian deposits in the Sentralbanken High and lowermost Norian samples from the Nordkapp Basin on the South Barents Shelf allow us to refine this picture, identify and explain temporal variations in source contributions, and discuss geographical variations in source.
Zircon age distributions in the lower strata (Ladinian) of the Snadd Formation, dominated by Uralian and Timanian orogen ages, indicate a provenance associated with the present-day Ural Mountains, where these rock ages are prevalent (Kuznetsov et al., 2007;Puchkov, 2009).However, at some point during the early Carnian, zircons with Triassic, Permian and Ordovician-Silurian ages were introduced alongside mainly Neoproterozoic and Palaeoproterozoic grains (Fig. 3).The region with rock ages and detritalzircon age populations that best fits with the zircon age spectra present in the Snadd and De Geerdalen formations is the Taimyr Peninsula in northern Siberia and surrounding areas (Letnikova et al., 2014;Ershova et al., 2016;Priyatkina et al., 2017).The 2500, 2000 Ma and Neoproterozoic populations have a clear association with Taimyr and North Siberia but are comparatively uncommon in Baltica and Greenland (Pózer Bue & Andresen, 2014;Ershova et al., 2016;Priyatkina et al., 2017).In Baltica and Laurentia, the Caledonian Orogeny produced zircon of early Palaeozoic age that typically co-occur with Mesoproterozoic zircon ages (Bingen & Solli, 2009).
The link to Taimyr is favoured based on the existence of several Permian to Middle Triassic granitoids related to the South Taimyr Orogen (Kurapov et al., 2021), and the documented Late Triassic volcanism in this region (Fig. 1; Letnikova et al., 2014).Detrital Cr-spinel in the Barents Basin lacks a LIP signature (Harstad et al., 2021), thus a Siberian craton source must be restricted to areas unaffected by Siberian Traps magmatism.It should be noted that the 425 Ma age peak observed in the detrital-zircon dataset remains enigmatic.Based on the available data, the Carnian/Norian strata of the Snadd and De Geerdalen formations appear to have been derived from the Taimyr region, i.e., in contrast to a dominantly Uralian source for the lower, Ladinian strata.The significance of this contrast is discussed further below.
The zircon Hf isotope data (Fig. 5) provide additional information relevant to distinguishing provenance.Of particular interest here are the youngest zircon grains, with ages close to the depositional age of the host sediments.The correspondence between detrital-zircon age and age of deposition can be considered an indication of an active, accretionary orogen source (Cawood et al., 2012), or melting of older continental crust by intrusive mafic magma at the onset of extension (Liu et al., 2019), consistent with the range in εHft values from negative (-4) to positive (+8), indicate a mix of juvenile and reworked crustal sources.As shown by Ershova et al. (2018), although Uralianderived detrital zircons also show considerable spread, they tend to plot at lower εHf values, consistent with the collisional nature of the Uralian Orogen (e.g., Spencer et al., 2019).Reworking of this Uralian crust in the Late Triassic would yield even more evolved Hf compositions, which is not observed in our data.Unfortunately, comparatively few Hf analyses were obtained from the older, Ladinian strata; thus, at the moment we are not able to document that the change in provenance, suggested by changing detrital-zircon ages with stratigraphic position, is also reflected in the Hf data.

Temporal change in source region rather than geographically distinct depositional centres with different sources
The gradual change in detrital-zircon age spectra with younging stratigraphic ages, as suggested by the data presented here, becomes even more prominent when combined with other published data from the Snadd and De Geerdalen formations (Fig. 6, see caption for data sources).Two age peaks (300 and 550 Ma) dominate in the lowest, Ladinian-lower Carnian deposits (Fig. 6D) and gradually decrease in relative size upward in the formations (Fig. 6A-C).In contrast, the 235 and 425 Ma detrital-zircon age peaks are absent or insignificant in the lowermost (Ladinian -lower Carnian) parts of the formations but become dominant in the uppermost (late Carnian -lowest Norian) deposits.
In deposits of uppermost Carnian to lowermost Norian age, the detrital-zircon age distributions in the Sverdrup Basin, on Franz Josef Land and in the Nordkapp Basin (Fig. 7) are dominated by one 235 Ma age peak, indicating the same detrital-zircon source to the entire area.The similarity in detritalzircon age distributions of sandstone samples located over 1000 km apart (present-day and likely palaeo distance), suggests a high degree of homogeneity in the detritus of this enormous depositional system.
It also favours the interpretation of a North Uralian source to the Pat Bay Formation in the eastern Sverdrup Basin (Omma et al., 2011;Hadlari et al., 2018).Considering the Late Triassic sediments of the eastern Sverdrup Basin as a continuation of the already enormous Snadd and De Geerdalen delta (Klausen et al., 2019;Gilmullina et al., 2022) significantly increases the size of the depostional systems.It also explains the Barents basin sediment spill-over documented by Gilmullina et al. (2021), as a continuation into the Sverdrup Basin.(Omma et al., 2011;Pózer Bue & Andresen, 2014;Soloviev et al., 2015;Fleming et al., 2016;Flowerdew et al., 2019;Khudoley et al., 2019;this study) Due to the northwestward migration of clinoforms in the Snadd -De Geerdalen delta (Glørstad-Clark et al., 2010;Klausen et al., 2015Klausen et al., , 2019;;Gilmullina et al., 2022), there is a rough correlation between geographical location and stratigraphic level within the Snadd Formation and equivalents.
When the hypothesis of geographical differences is tested by investigating the top Snadd deposits from a southern location in the Nordkapp Basin, we produced a detrital-zircon age pattern similar to top Snadd equivalents in the far north of the basin.Thus, a hypothesis of consistent geographical differences in zircon age spectra is not supported by the new data (Fig. 7), nor is it supported by previously published detrital Cr-spinel data (Harstad et al., 2021).
Considering the data presented in this and other publications, intra-formational stratigraphic variation and gradual source-terrain evolution/shift represents an alternative interpretation for the observed variations in zircon age spectra, suggesting that temporal changes in source rather than geographical variations are a more likely explanation.Figure 8 shows the evolution in provenance of the Snadd and De Geerdalen formations, as proposed in this contribution.The model involves changes in sediment contribution to relatively stable sediment pathways and is based on the new data presented herein and considerations of the Arctic regional geological evolution, summarised above.The depositional system entered the Barents Basin in the southeast and transported sediments in a northwestward direction, as determined from seismicity-based research (e.g., Klausen et al., 2015).In the Ladinian and early Carnian, the system deposited sediments from the North Uralian and Timanian orogens (Fig. 8A), consistent with a lack of a significant 235 Ma detrital zircon peak.In the early Carnian, however, a younger, 235 Ma component heralds the contribution of a new sediment source, mixed with the Timanian and Uralian orogen sediments.The new source is likely to be in the Taimyr Peninsula (Fig. 8B), where volcanism of this age is known (Letnikova et al., 2014;Kurapov et al., 2021;and references therein).The tectonic explanation for this volcanism, and/or tectonism with concomitant production of considerable sediment volumes in Taimyr in this period cannot be answered based on the data provided in this study.In the middle to late Carnian and earliest Norian, the Taimyr source contributed a progressively larger proportion of detritus (Fig. 8C).The continued contribution from the Uralian Orogen, observed in both detrital zircon and Cr-spinel datasets (Harstad et al., 2021),  (Omma et al., 2011;Hadlari et al., 2018), the Northeast Barents Shelf (Soloviev et al., 2015)   suggests that the original supply system did not terminate, but with time grew less significant in terms of sediment delivered (Fig. 8D).It is possible that the youngest zircon grains in the uppermost parts of the formations entered the basin as volcanic ash, expressed as the ~235 Ma age peak.Such an interpretation is supported by the presence of abundant volcanic clasts in the sediments (e.g., Mørk, 1999).
The observed increase in two different age-peaks, the 235 Ma and the 425 Ma peaks, combined with the long distance to any known, active volcanic region at the time (the Taimyr Peninsula), favours a significant water-transported contribution to the deposits.
We suggest two analogues to the sedimentary environment discussed above: the present-day Ganges and Brahmaputra drainage system, or the paleo-Mississippi drainage system during the Cenozoic (e.g., Craddock et al., 2013;Blum et al., 2017).In the Ganges-Brahmaputra analogue, Brahmaputra contributes sediment from behind the highest peaks of the Himalayas, while the Ganges represents mountain chain-parallel transport of sediments.We suggest the interpreted Taimyr source to be like a Brahmaputra that steadily increased its sediment contribution to the combined system.The Ganges, on the other hand, is suggested as an analogue to the river draining the remnants of the Uralian and Timanian orogens.Studies of Cenozoic sediments in the Gulf-of Mexico (Craddock et al., 2013;Blum et al., 2017) 2014), and the other a Uralian source, similar to that of the Snadd and De Geerdalen formations, farther southeast.Westerly derived sediments have been documented in this area previously (Mørk et al., 1982;Glørstad-Clark et al., 2010), while Harstad et al. (2021) concluded that the Cr-spinel composition in these sediments is similar to that of the Snadd and De Geerdalen formation samples, suggesting little mafic input from western Greenland sources.

Conclusions
The detrital-zircon U-Pb and Lu-Hf data presented here from Ladinian to Norian deposits of the Snadd and De Geerdalen formations on the Barents Shelf are consistent with previous studies arguing for Uralian and Timanian sources.In contrast to previous studies, we show that there is a basin-wide temporal change in provenance, consistent with an initial Uralian source that was gradually replaced by another, Taimyr Peninsula source, which produced zircon similar to the age of deposition.
The prograding depositional system likely reached the Sverdrup Basin in the middle or late Carnian while in West Svalbard, a second minor source region, likely Greenland, is documented in a mixed Greenland and Uralian/Taimyr detrital zircon population.
This study shows the potential for correlation of units over vast distances on the Barents Shelf, identification of particular stratigraphic levels, and acquiring information about the tectonic evolution of source regions reflected in temporally varying sedimentary input.

Figure 1 .
Figure 1.Palaeogeographic reconstruction of the polar areas in the Late Triassic showing areas of several major tectonic events of the Neoproterozoic to Late Triassic, major areas of Triassic deposition and regional locations.As the figure represents an extended period, the considerable overlap between units is not accounted for in the figure.Late Triassic palaeogeography, extent of depositional and offshore areas modified afterSømme et al. (2018).Outline of the Timanian and Caledonian Orogens afterGee et al. (2006),Willner et al. (2019) andErshova et al. (2016), outline of the present-day Ural mountains afterPuchkov (2009), outline of the Central Asian Orogenic Belt afterXiao et al. (2015) andErshova et al. (2016), outline of the Siberian Traps large igneous province afterIvanov et al. (2018) and the outline of the Kara Orogen afterVernikovsky et al. (2020).The South Taimyr event is suggested based onZhang et al. (2018a).Possible location of Carnian volcanic activity based onLetnikova et al. (2014) and references therein.
Flowerdew et al. (2019) discussed the possibility of a potential mixed source, with recently eroded material and recycled Uralian Orogen detritus, that accounts for apparent similarities between these sources.Other single-grain chemical and/or geochronological methods have generally supported a Uralian Orogen provenance for the Snadd and De Geerdalen formations.Flowerdew et al. (2019) describe detrital-apatite age distributions similar to the detrital-zircon age distributions.Harstad et al.
Figure 2. (A) Map of the Norwegian Barents Shelf with outlines of structural highs and basins, modified from Lundschien et al., 2014.Sample locations are indicated with black spots and numbered as in Table1.(B) Lithostratigraphic overview of the Triassic of the sampled area(Lundschien et al., 2014).
. Ladinian and uppermost Carnian deposits are sampled from shallow cores in the Nordkapp Basin, lower Carnian at the Sentralbanken High, and middle Carnian deposits in the Bjarmeland Platform area.The samples collected from outcrops on Svalbard represent upper Carnian and possibly lowermost Norian deposits.
, the Ladinian deposits, irrespective of location, are compared to Carnian -early Norian deposits and sample 11 from West Spitsbergen.Mesoproterozoic ages dominate the pre-635 Ma zircon populations in the Ladinian samples and sample 11.The pre-635 Ma zircon distribution in the Carnian to lower Norian samples contains fewer Mesoproterozoic zircon grains and has a higher proportion of Neoproterozoic grains, as well as the presence of 2000 and 2500 Ma peaks.

Figure 3 .
Figure 3. Zircon age distribution diagrams for all samples analysed in this study.The samples are ordered stratigraphically, with the Ladinian samples lowest and early Norian samples at the top.The lowermost samples are dominated by detrital zircons with ages corresponding to the Uralian and Timanian orogens.This distribution is inverted upwards, with just a minor proportion of Carboniferous -Permian grains, and a dominance of Triassic grains in the youngest strata.The only sample with a sizable proportion of Mesoproterozoic and older zircon is located in West Svalbard, indicating a mixed provenance at this location.Arrows denote depositional age.

Figure 4 .
Figure 4. Density-distribution diagrams for the 3000-600 Ma populations of the Ladinian, Carnian -early Norian and West Spitsbergen detrital zircons of this dataset.The Precambrian zircons from West Spitsbergen display patterns similar to earlier Triassic samples (Pózer Bue & Andresen, 2014), and show a mixed provenance area.The Ladinian deposits contain mostly Late Palaeoproterozoic and Mesoproterozoic zircons, but few early Palaeoproterozoic and Neoproterozoic grains.In the Carnian-early Norian deposits, the Mesoproterozoic is less pronounced while Early Palaeoproterozoic and Neoproterozoic zircon ages are prevalent.The observed differences match with observed ages of the Baltica and Siberia cratons, with late Palaeoproterozoic and Mesoproterozoic rocks associated with Baltica (Pózer Bue & Andresen, 2014), and early Palaeoproterozoic and Neoproterozoic rocks found in and around Siberia(Ershova et al., 2016;Priyatkina et al., 2017).However, care must be taken when attempting to identify particular continents based solely on detrital-zircon data(Andersen et al., 2016;Slagstad & Kirkland, 2017)

Fleming
et al. (2016) were the first to discuss the zircon age variations within the Snadd and De Geerdalen formations and their Franz Joseph Land equivalents.They attributed the observed variations to a difference between the northern and the southern parts of the Uralian Orogen and suggested that the Triassic and Ordovician -Silurian zircons observed in the north better represent Taimyr or north Siberia.The southern samples, on the other hand, were dominated by Devonian -Carboniferous and Ediacaran -Cambrian detrital zircons that can be directly associated with a source in the Uralian and Timanian orogens.

Figure 6 .
Figure6.Compilation of all relevant published detrital zircon ages in, or time-equivalent with, the Snadd and De Geerdalen formations, ordered by time of deposition(Omma et al., 2011; Pózer Bue & Andresen, 2014;Soloviev et al., 2015;Fleming et al., 2016;Flowerdew et al., 2019; Khudoley et al., 2019; this study).Complete relative density plots for the time period 3000 -200 Ma to the left and a focus on the 600 -200 Ma distributions to the right.We observe a gradual change from a dominant 400 -300 Ma zircon population in the lowest deposits, through introduction of Triassic and 480 -400 Ma zircon grains, with an ultimate dominance of Triassic zircons.The uppermost unit is a composite of samples from the Sverdrup Basin, South Barents Sea and Franz Joseph Land.
Figure6.Compilation of all relevant published detrital zircon ages in, or time-equivalent with, the Snadd and De Geerdalen formations, ordered by time of deposition(Omma et al., 2011; Pózer Bue & Andresen, 2014;Soloviev et al., 2015;Fleming et al., 2016;Flowerdew et al., 2019; Khudoley et al., 2019; this study).Complete relative density plots for the time period 3000 -200 Ma to the left and a focus on the 600 -200 Ma distributions to the right.We observe a gradual change from a dominant 400 -300 Ma zircon population in the lowest deposits, through introduction of Triassic and 480 -400 Ma zircon grains, with an ultimate dominance of Triassic zircons.The uppermost unit is a composite of samples from the Sverdrup Basin, South Barents Sea and Franz Joseph Land.

Figure 7 .
Figure7.A comparison of detrital-zircon age distributions in samples from the Sverdrup Basin(Omma et al., 2011;Hadlari et al., 2018), the Northeast Barents Shelf(Soloviev et al., 2015) and the Southern Barents Shelf (this study).Samples are of latest Carnian and early Norian age.The similar zircon age distributions demonstrate a similar source to all samples.The eastern parts of the Pat Bay Formation in the Sverdrup Basin can therefore be considered a continuation of the Snadd and De Geerdalen depositional system.

Figure 8 .
Figure 8. Provenance interpretation and suggested sediment paths into the Barents and Sverdrup palaeobasins for four consecutive Middle -Late Triassic time intervals.(A) The Ladinian and early Carnian detrital-zircon age distributions contain two main age peaks at 300 Ma and 540 Ma, indicating a provenance from the Uralian and Timanian orogens in the north Urals.(B-D) Higher proportions of other detrital zircon populations with ages of 235 and 425 Ma, and reduction in 300 Ma detrital zircon population in the early Carnian to the early Norian suggest gradually increasing input from another source region, likely the Taimyr Peninsula.(D) In late Carnian, a mixed provenance of Greenland and Ural/Taimyr Peninsula is present in West Spitsbergen.The maximum extent of the depositional system was reached in latest Carnian/early Norian and likely extended to include the East Sverdrup Basin.The map is modified from Sømme et al. (2018), with the distribution of the Snadd and De Geerdalen depositional system taken fromKlausen et al. (2015) andAnfinson et al. (2016).
reveal variations in the relative abundance of detrital zircon ages, interpreted to represent variations in the catchment-area of the paleo-Mississippi, a development highly comparable to what we propose for the Snadd and De Geerdalen drainage system.Early Norian development on the basin fringes Sample 11, from the late Carnian/early Norian De Geerdalen formation in Kapp Toscana, West Spitsbergen, has a different detrital-zircon age distribution compared with the other samples in this study.Although the West-Spitsbergen sample has a 500-200 Ma zircon age distribution that is similar to other De Geerdalen-formation samples, the 3000-500 Ma distribution is similar to Early and Middle Triassic sediments derived from the west, most likely from Greenland (Pózer Bue & Andresen, 2014).This distribution is likely the result of mixing of two sediment sources, one representing a continuation of the Middle Triassic Greenland-derived depositional system (Pózer Bue & Andresen, Our new data suggest a temporal change in source region associated with a gradual change in dominant zircon age peaks from Ladinian (bottom) to Norian (top) deposits.With the new data, top Snadd Formation zircon age distributions are shown to contain remarkably similar detrital-zircon age distributions over a vast area from the Nordkapp Basin in the south Barents Sea to the eastern Sverdrup Basin.A new local source to western Svalbard in the late Carnian is also documented.The source to the detrital-zircon grains in the Ladinian and lower Carnian deposits of the Snadd Formation is likely Uralian and Timanian orogen rocks in the present-day North Urals.The other progressively more important detrital-zircon source is likely to have been the palaeo-Taimyr area undergoing active tectonism and volcanism at the time.In the Snadd and De Geerdalen formations, we observe a gradual change in detrital-zircon age spectra distribution.This change is characterised by systematic variation in detrital-zircon age-peak distributions.In the oldest deposits, there are two dominant age peaks at 300 Ma and 540 Ma, that gradually change into one dominant 235 Ma age peak and two minor age peaks at 300 Ma and 425 Ma in the youngest deposits.We interpret the change in zircon age distribution to be caused by the introduction and progressive increase of a new, and eventually dominant, zircon source region.

Table 1 .
An overview of studied samples including number of analysed zircon grains.Sample 1-8 are Snadd Fm. samples, sample 9-11 are from the De Geerdalen Fm.
Sample numbers are arranged geographically; Southeast (1) to Northwest (11).Stratigraphic age based on biostratigraphic ages in Triassic peak (purple colour) increases in significance stratigraphically upward, from an insignificant peak in the oldest units to a dominant peak in the youngest.Late Devonian to Permian zircon (brown, pale blue and red), corresponding to the age of the Uralian Orogen, is common in all samples but the distribution of ages varies in detail.In the oldest, Ladinian samples, where this age group is the proportionally largest, the predominant age peak is 360 Ma, with Permian ages almost absent.In upper Carnian and lower Norian samples, the proportion of Late Devonian to Permian zircon remains high, although there are fewer Devonian and more Permian zircon grains in the youngest deposits.