Rare earth elements and neodymium and strontium isotopic constraints on provenance switch and post-depositional alteration of fossiliferous Ediacaran and lowermost Cambrian strata from Arctic Norway

Peninsula in northeastern Finnmark, Arctic Norway, comprises one of the most complete Ediacaran – Cambrian transitions worldwide with a nearly continuous record of micro-and macrofossils from the interval of the diversification of complex life. Here, we report on the provenance and post-depositional alteration of argillaceous mudstones from the Digermulen Peninsula using rare earth elements and Sm – Nd and Rb – Sr isotopic systematics to provide an environmental context and better understand this important transition in Earth ’ s history. The studied sections comprise a mid-Ediacaran glacial – interglacial cycle, including the Nyborg Formation (ca. 590 Ma) and Mortensnes Formation (related to the ca. 580 Ma-old Gaskiers glaciation), and the St ´ ahpogieddi Formation (ca. 560 – 537 Ma), which yields Ediacara-type fossils in the Indreelva Member and contains the Ediacaran – Cambrian boundary interval in the Manndrapselva Member and basal part of the informal Lower Breidvika member (ca. 537 – 530 Ma). Three sample groups, (1) Nyborg and Mortensnes formations, (2) the lowermost five samples from the Indreelva Member and (3) the remaining samples from the Indreelva as well as from the Manndrapselva and Lower Breidvika members, can be distinguished, belonging to distinct depositional units. All samples have negative ε Nd (T) values ( (cid:0) 6.00 to (cid:0) 21.04) indicating a dominant input of terrigenous detritus with an old continental crust affinity. Significant shifts in Sm – Nd isotope values are related to changes in the sediment source, i.e. Svecofennian province vs Karelian province vs Svecofennian province plus in addition likely some juvenile (late Neoproterozoic volcanic) material, and probably reflect palaeotectonic reorganisation along the Iapetus-facing margin of Baltica. The combined Rb – Sr isotopic data of all samples yield an errorchron age of about 430 Ma reflecting the resetting of the Rb – Sr whole-rock isotope systems of the mudstones during the Scandian tectono-metamorphic event in the Gaissa Nappe Complex of Finnmark.

The Digermulen Peninsula in northeastern Finnmark, Arctic Norway, comprises one of the most complete Ediacaran-Cambrian transitions worldwide with a nearly continuous record of micro-and macrofossils from the interval of the diversification of complex life.Here, we report on the provenance and post-depositional alteration of argillaceous mudstones from the Digermulen Peninsula using rare earth elements and Sm-Nd and Rb-Sr isotopic systematics to provide an environmental context and better understand this important transition in Earth's history.The studied sections comprise a mid-Ediacaran glacial-interglacial cycle, including the Nyborg Formation (ca.590 Ma) and Mortensnes Formation (related to the ca.580 Ma-old Gaskiers glaciation), and the Stáhpogieddi Formation (ca.560-537 Ma), which yields Ediacara-type fossils in the Indreelva Member and contains the Ediacaran-Cambrian boundary interval in the Manndrapselva Member and basal part of the informal Lower Breidvika member .Three sample groups, (1) Nyborg and Mortensnes formations, (2) the lowermost five samples from the Indreelva Member and (3) the remaining samples from the Indreelva as well as from the Manndrapselva and Lower Breidvika members, can be distinguished, belonging to distinct depositional units.All samples have negative ε Nd (T) values (− 6.00 to − 21.04) indicating a dominant input of terrigenous detritus with an old continental crust affinity.Significant shifts in Sm-Nd isotope values are related to changes in the sediment source, i.e.Svecofennian province vs Karelian province vs Svecofennian province plus in addition likely some juvenile (late Neoproterozoic volcanic) material, and probably reflect palaeotectonic reorganisation along the Iapetus-facing margin of Baltica.The combined Rb-Sr isotopic data of all samples yield an errorchron age of about 430 Ma reflecting the resetting of the Rb-Sr whole-rock isotope systems of the mudstones during the Scandian tectono-metamorphic event in the Gaissa Nappe Complex of Finnmark.Preservation of palaeopascichnids coincides with the sedimentation regimes of sample groups 2 and 3 while other Ediacara-type fossils, e.g.Aspidella-type and frondose forms, are limited to the sample group 3. Our results are similar to those of earlier studies from the East European Platform in suggesting oxic seafloor conditions during the late Ediacaran.

Introduction
The Ediacaran Period is an important interval of Earth's history when metazoans started to flourish following the 'Snowball Earth' glaciations (Cunningham et al., 2017;Wood et al., 2019, and references therein).To better understand the depositional environment that may have benefited or hindered the evolution of this early complex life, the Ediacaran sedimentary successions are studied in great detail worldwide, through whole-rock geochemical and isotopic studies, among other proxies.Geochemical and isotopic studies of siliciclastic sedimentary rocks are critical for deciphering the age and composition of rock lithologies in the source area and hence for palaeotectonic reconstructions.Fine-grained siliciclastic sediment (mud) or sedimentary rock (mudstones, according to the nomenclature of Lazar et al., 2015) is thought to be the most appropriate lithology for such studies because the detritus is thoroughly mixed during weathering processes and transport of sediments, and chemical modification of the sediment during deposition and burial diagenesis is less pronounced than in coarser sediment (Allègre and Rousseau, 1984;Condie, 1991;Cullers, 1995).Thus, they provide a representative sampling of the provenance, i.e. the exposed upper continental crustal rocks at the time of deposition of the sediment (Taylor and McLennan, 1985).
Rare earth elements (REEs), also called lanthanides, are widely used as tracers for redox changes and sedimentary provenance of fine-grained sedimentary rocks (e.g., McLennan, 1989;McLennan et al., 1990McLennan et al., , 1993)).The REEs are a group of chemically coherent elements, comprising lanthanum (La) to lutetium (Lu), that all exist in a trivalent oxidation state except for cerium (Ce) and europium (Eu) which are redox sensitive and may take different valence states depending on the oxygen fugacity (fO 2 ) (Elderfield and Greaves, 1982).At high fO 2 , Ce can be in the tetravalent oxidation state (Ce 4+ ), while under low fO 2 , Eu can be in the divalent oxidation state (Eu 2+ ) (Elderfield and Greaves, 1982;Gao and Wedepohl, 1995).Besides the REEs, also Sm-Nd whole-rock isotope systematics have proved to be a valuable tool for unravelling the provenance of fine-grained sedimentary rocks (McCulloch and Wasserburg, 1978;O'Nions et al., 1983;Allègre and Rousseau, 1984;Goldstein et al., 1984;Nelson and DePaolo, 1988;McLennan and Hemming, 1992;Bock et al., 1994;Roy et al., 2007;Awasthi et al., 2018).This is because Sm/ Nd fractionation occurs largely during extraction of melts from the mantle to the crust (McCulloch and Wasserburg, 1978;DePaolo, 1981), whereas processes such as erosion, weathering, transport, sedimentation, diagenesis and metamorphism do not fractionate Sm from Nd appreciably.Both elements belong to the light REE (LREE include La, Ce, Pr, Nd, Pm, and Sm), and thus the Nd isotope signature of the detritus carried into sedimentary basins reflects the average lithology and average 'crustal residence age' of the eroded source rocks (e.g., McCulloch and Wasserburg, 1978;Allègre and Rousseau, 1984;Goldstein et al., 1984;McLennan and Hemming, 1992;Roy et al., 2007;Awasthi et al., 2018).Changes in the isotope composition and crustal residence ages throughout a stratigraphic section are utilized to unravel prominent changes in provenance, reflecting continent-scale tectonic reorganisation.Crustal residence ages are often referred to as model ages (Arndt and Goldstein, 1987).It should be noted that the Nd isotope system may be reset due to mobility of LREE under certain circumstances during diagenesis (Awwiller and Mack, 1991;Milodowski and Zalasiewicz, 1991;Bock et al., 1994Bock et al., , 1998)).Regardless of this observation, it is still possible to extract important provenance information from Sm-Nd isotopic data (Bock et al., 1998). 87Rb-87 Sr whole-rock isotopic data are also informative to study past depositional settings, although the data are less reliable for provenance studies because the Sr isotope values (e.g., 87 Sr/ 86 Sr) can be modified during fluvial transport (Awasthi et al., 2018) and diagenesis (McCulloch and Wasserburg, 1978;Ohr et al., 1991).Apart from that, already low-grade metamorphism commonly resets the 87 Rb-87 Sr whole-rock isotope system, and thus, the 87 Rb-87 Sr whole-rock isotope method is useful to determine the age of metamorphic overprint (Compston and Pidgeon, 1962;Gebauer and Grünenfelder, 1974) but is inadequate for finding the composition and age of the source area of metasedimentary rocks. 87Rb-87 Sr isotopic analysis were carried out on upper Neoproterozoic sedimentary rocks from the Varanger Peninsula in Finnmark (Pringle, 1973;Gorokhov et al., 2001) but have not yet been performed on sedimentary rocks from the adjacent Digermulen Peninsula to the west (Figs. 1, 2).
Here, we apply for the first time 147 Sm-143 Nd and 87 Rb-87 Sr wholerock isotopic analyses to low-grade metamorphosed mudstones from the Digermulen Peninsula in northern Norway (Figs. 1,2), in order to understand sediment composition, provenance and post-depositional alteration.This area is the only fossiliferous site in Scandinavia with sedimentation across the Ediacaran-Cambrian transition without a significant hiatus (Högström et al., 2013) (Fig. 3).Furthermore, it is the only locality in Scandinavia where Ediacara biota fossils have been found (Farmer et al., 1992;Högström et al., 2013;Jensen et al., 2018a,b) (Fig. 4), and a goal of this work is to understand the way these fossils are preserved as well as the conditions they were deposited in.Our 147 Sm-143 Nd isotopic data provide the first constraints on provenance changes across a glacial-interglacial cycle and the fossil-bearing Ediacaran-Cambrian boundary interval in Scandinavia.Results from the 87 Rb-87 Sr whole-rock isochron method (this study) are compared with previously obtained ages for the metamorphic overprint based on finefraction 40 K-40 Ar (Meinhold et al., 2019b) and in situ 238,235 U-206,207 Pb calcite data (Meinhold et al., 2020).Finally, we explore how redox-sensitive elements (primarily Ce) in the Digermulen succession relate to the stratigraphical distribution of palaeopascichnids and Ediacara-type fossils.The appearance of probable metazoan macroscopic fossils of the Ediacara biota at approximately 575 Ma, as well as later Ediacaran diversity changes and extinctions, have been linked to the late Neoproterozoic rise in seawater oxygenation and to Ediacaran anoxic events (e.g., Sahoo et al., 2016;Evans et al., 2018;Zhang et al., 2019;Rooney et al., 2020;Shi et al., 2022).The data from the continuous Digermulen succession cast further light on this discussion.

Regional geology
The sedimentary succession of the Digermulen Peninsula belongs to the Lower Allochthon Gaissa Nappe Complex of the Caledonides of Finnmark (Rice, 2014) (Fig. 1).The rocks belong to the Cryogenian to lower Cambrian Vestertana Group, comprising in ascending order the Smalfjorden, Nyborg, Mortensnes, Stáhpogieddi and Breidvika formations (Reading, 1965;Reading and Walker, 1966;Banks et al., 1971) (Fig. 2, 3).The glacial deposits of the Smalfjorden and Mortensnes formations (e.g., Reading and Walker, 1966;Banks et al., 1971), separated by the interglacial Nyborg Formation (Reading and Walker, 1966), are collectively known as the Varanger Ice Age (e.g., Nystuen, 1985).The Smalfjorden Formation has been correlated with glaciogenic deposits of the Marinoan glaciation (e.g., Rice et al., 2011) of Cryogenian age (ca. 645-635 Ma, Rooney et al., 2015;Shields-Zhou et al., 2016).The Mortensnes Formation on the other hand has been correlated with those of the Ediacaran Gaskiers glaciation (e.g., Rice et al., 2011), which was likely of short duration (ca.580 Ma, Pu et al., 2016).The age of these glacial deposits has recently been challenged by Kumpulainen et al. (2021), who suggest that all Neoproterozoic glacial successions in Scandinavia probably belong to the Marinoan glaciation, which is a global Snowball Earth glacial episode.Kumpulainen et al. (2021) base their interpretation on cross-cutting relationships between glacial successions and a ca.596 Ma-old mafic igneous dyke in the Särv Nappe of the central Scandinavian Caledonides in Jämtland, west-central Sweden.A Marinoan age may be true for the study area in Jämtland.However, as long as no reliable and unequivocal geochronological data are available from Finnmark and independent proof that the age constraint from  The postglacial Stáhpogieddi Formation is divided into the Lillevannet, Indreelva and Manndrapselva members (Fig. 3).The lower part of the formation represents a transgressive sequence from the fluvial conditions of the upper part of the Lillevannet Member into the quiet marine conditions of the Indreelva Member (Banks et al., 1971;Banks, 1973), with deposition in a marine environment continuing into the Breidvika Formation (Reading, 1965;Banks et al., 1971).The Stáhpogieddi and Breidvika formations mainly comprise siliciclastic sedimentary rocks (mudstones, siltstones, and sandstones).Recently, Meinhold et al. (2019a) described carbonate beds, lenses and concretions from the 2nd cycle of the Manndrapselva Member.In a simplified view, palaeocurrent flow directions are mainly to the northwest and north in the older strata (Nyborg and Mortensnes formations and Lillevannet Member), to the east in the Indreelva Member, and to the southwest in the younger strata, with the informal Lower Breidvika member showing bimodal directions (Fig. 3).
On the Digermulen Peninsula, Ediacara-type fossils occur in the Indreelva Member (Farmer et al., 1992;Högström et al., 2013;McIlroy and Brasier, 2017;Jensen et al., 2018a,b) (Fig. 3).Aspidella-type discoidal fossils first appear close to the base of the Indreelva Member (Figs. 3,4a), within a facies dominated by red and green siltstones.A more diverse association of Ediacara-type fossils are found higher in the Indreelva Member in a facies with a higher proportion of sandstone beds.This includes frondose forms (Fig. 4b) and a dickinsoniomorph.In addition, flask-shaped microfossils have recently been described from Fig. 3. Simplified stratigraphy of the Vestertana Group (after Jensen et al., 2018b).The distribution of Sabellidites cambriensis is according to Ebbestad et al. (2022).Palaeocurrent data were taken from the published literature (Reading, 1965;Banks et al., 1971;Banks, 1973).As an exemplary explanation, an arrow pointing to the top indicates a palaeocurrent flow direction from south to north.On the right side, the stratigraphic range and lithological logs of the two sections analysed in this study are shown with stratigraphic positions of mudstone samples taken for chemical and isotopic analyses being indicated.Stratigraphic heights in Sections 1 and 2 are given with reference to the top of the Lillevannet Member and the Indreelva Member, respectively.The grey and red-brown colours next to the stratigraphic heights indicate the approximate colours of the rock strata.The stratigraphic position of the boundary between the sample groups P2 and P3 is somewhere between samples D14-AP56 and D14-AP56a.The stratigraphic positions of the two mudstone samples from the Nyborg Formation (D14-AP32 and D14-AP32.1)and the single sample from the Mortensnes Formation (D14-AP33) are indicated in the simplified stratigraphic column on the left side.

Detrital zircon U-Pb data
For the later discussion of the sediment provenance based on isotopic data (Section 5.3.) it is essential to summarize some of the published data about detrital zircon U-Pb ages from the Ediacaran and Cambrian sandstones of Finnmark.Detrital zircon U-Pb ages from a sandstone sample of the interglacial Nyborg Formation from the adjacent Varanger Peninsula are mainly 2.0-1.85Ga, with a minor spread between 3.0 and 2.5 Ga (Roberts and Siedlecka, 2012).No Mesoproterozoic zircon grains were detected.The omnipresence of mid-Palaeoproterozoic zircon grains in the Nyborg Formation points toward a sediment source from the Svecofennian province, with only a minor influx from the Karelian province.
Detrital zircon U-Pb ages from a Manndrapselva Member sandstone sample collected on the mainland of northern Finnmark, just south of the Digermulen Peninsula, yielded one major age peak at 554 Ma, a broad spread of ages between 1620 and 930 Ma, and a subordinate age peak at 645 Ma (Zhang et al., 2015).Also, lower Cambrian sandstones from the Dividalen Group close to the international border between Finland and Norway contain detrital zircon grains with Ediacaran ages besides a broad spread of ages between ca.2000 and 1000 Ma and a few late Archaean grains (Andresen et al., 2014).Overall, the detrital zircon age spectra from sandstones of the Manndrapselva Member and the Dividalen Group are evidence for mixed sediment influx from old upper crustal sources from the Svecofennian province and juvenile detrital material (ca.570-550 Ma); the latter is interpreted to be derived from magmatic rocks of the Timanides (Andresen et al., 2014;Zhang et al., 2015).This orogen formed during the late Ediacaran along the northeastern periphery of Baltica (present-day coordinates) due to accretion of island-arc complexes onto Baltica (Pease et al., 2004).

Material and methods
This study is based on fine-grained siliciclastic sedimentary rock samples from surface outcrops along two stratigraphic sections at the eastern side of the Digermulen Peninsula (Figs. 2, 3).In order to understand sediment composition, provenance and post-depositional alteration a total of 21 samples were processed for bulk mineralogy, REE chemistry, and 147 Sm-143 Nd and 87 Rb-87 Sr isotopic analyses.Stratigraphic levels of the samples are shown in Fig. 3.Most of the mudstones are greenish-grey, a few are bluish-grey in colour.Sample D14-G6 is the only sample of brownish-grey colour.We chose finegrained siliciclastic sedimentary rocks because they normally best reflect the average composition of the source area compared to sandstones or coarser sediments (McCulloch and Wasserburg, 1978;Allègre and Rousseau, 1984;Cullers, 1995).The majority of the samples were also processed for palynological analysis.Initial results (Palacios et al., 2017;Agić et al., 2019Agić et al., , 2022;;Teodoro Palacios, unpublished data) as well as biostratigraphic information from macrofossils (Högström et al., 2013;McIlroy and Brasier, 2017;Jensen et al., 2018a,b;Ebbestad et al., 2022) provide constraints on the depositional ages.
The dried sample material was disintegrated with a hammer and thoroughly powdered and homogenized using an agate ball mill.Some of the rock powder was used for X-ray diffraction (XRD) analysis to determine semi-quantitatively the bulk mineralogy of the samples.The analysis was conducted using a Philips X'Pert MPD, equipped with a PW3050 Goniometer.CuKα-radiation, a tube voltage of 40 kV, a tube current of 30 mA, and a scanning range from 4 • to 69.5 • 2θ was applied.Mineral identification was performed with X'Pert HighScore software package.Semi-quantitative data is based on the RIR method of X'Pert HighScore.XRD analysis was conducted at the Department of Sedimentology and Environmental Geology at the University of Göttingen, Germany.
The remaining rock powder was used for chemical and isotopic analyses.REE were quantified using inductively coupled plasma mass spectrometry (ICPMS) on the dissolved sample material.For each sample, 100 mg of rock powder were digested in the following steps using a PicoTrace® digestion system: (1) pre-reaction with 2 ml HNO 3 at 50 • C overnight, (2) first pressure phase with additional 3 ml HF (40%) and ml HClO 4 (70%) at 150 • C for 8 h, (3) evaporation at 180 • C for 16 h, (4) second pressure phase with 10 ml double de-ionised water, 2 ml HNO and 0.5 ml HCl at 150 • C for 4 h.The resulting solution was diluted to 100 ml with ultrapure water.Analysis was performed using a Thermo-Electron VG PlasmaQuad 2 quadrupole ICPMS at the Department of Geochemistry and Isotope Geology at the University of Göttingen.The REE data were normalized to both CI chondrite and PAAS (Post-Archaean Australian Shale) using the values of Pourmand et al. (2012).Normalisation of trace concentration data removes the natural 'evenodd' variability in the relative cosmological abundances of elements in solar system objects.It thus produces 'smooth' patterns of the samples in trace element plots and allows a more convenient comparison of their patterns.
The calculation of the Ce and Eu anomalies is common practice when dealing with whole-rock geochemical data of fine-grained sedimentary rocks as the Ce anomaly may provide information on redox conditions at the time of deposition (e.g., Wilde et al., 1996;Pattan et al., 2005) and the Eu anomaly on the provenance of the sediment, e.g.composition of the source area (e.g., McLennan, 1989;McLennan et al., 1993;Gao and Wedepohl, 1995).Ce anomalies are increasingly used as a proxy for Earth's oxygenation events and geobiological changes during the Proterozoic and early Phanerozoic (Wallace et al. 2017).Ce anomalies (Ce/ Ce*) were calculated following Bau and Dulski (1996): Ce/Ce* = Ce N / (0.5La N + 0.5Pr N ), where the subscript N indicates PAAS-normalized values.To test whether the obtained Ce anomaly values are genuine and not an effect of anomalous La enrichment, we applied the method of Bau and Dulski (1996) and additionally calculated the Pr anomalies (Pr/ Pr*) as follows: Pr/Pr* = Pr N /(0.5CeN + 0.5Nd N ).A negative Ce anomaly inevitably results in Pr/Pr* > 1, whereas a positive Ce anomaly generates Pr/Pr* < 1.The combination of Ce/Ce* < 1 and Pr/Pr* < 1, however, indicates a positive La anomaly and hence non-redox related Ce anomalies (Bau and Dulski, 1996).The Eu anomalies (Eu/Eu*) were calculated as suggested by McLennan (1989): Eu/Eu* = Eu N /(Sm N × Gd N ) 0.5 , where the subscript N indicates chondrite-normalized values.
Multicollector (MC)-ICPMS analyses of 143 Nd/ 144 Nd and 87 Sr/ 86 Sr isotope compositions were performed at the Department of Geochemistry and Isotope Geology at the University of Göttingen.The wetchemical procedures were performed in class 100 fume hoods housed in a metal-free, class 10,000 clean room environment.Samples were processed with double-distilled HCl, HNO 3 , and HF prepared by subboiling and diluted with high-purity water (18.2MΩ cm − 1 ) for chromatography.About 50 to 100 mg of sample powder (equivalent to ca. 300 ng Nd and 10 μg of Sr) were weighted into 15 ml screw top PFA beakers and spiked with 150 Nd-149 Sm and 87 Rb-84 Sr tracer solutions.Sample powders were digested in a mixture of 2 ml concentrated HNO 3 and 2 ml concentrated HF on a hotplate at 120 • C for 72 h and heated at 80 • C to incipient dryness.The sample cake was dissolved in 10 ml of 6 M HCl and heated at 110 • C overnight for complete sample dissolution.When necessary, this step was repeated.
The separation of Sr and Nd from the sample solution generally followed the method described in Hegner et al. (1995Hegner et al. ( , 2010)).The cation exchange procedures employed BioRad AG 50 × 8 (200-400 mesh) resin and Triskem Ln-Resin (100-150 μm) to separate Nd and Sm from neighbouring REE.The Sr cut was further purified with Triskem Sr-Resin (50-100 μm; method of Pin et al., 2014) for elimination of any remaining 87 Rb interfering with 87 Sr during mass spectrometry.The element fractions were diluted in 0.5 M HNO 3 to a nominal concentration of samples and reference materials of 100 ng mL − 1 for mass spectrometry.
The isotope ratios were determined on a ThermoFisher Scientific Neptune Plus MC-ICPMS equipped with nine Faraday cups.The ion currents of Sr and Nd isotopes were measured with 10 11 Ω amplifiers and the interfering isotope monitors (i.e., 140 Ce and 149 Sm for Nd isotopes, 83 Kr and 85 Rb for Sr isotopes, and 145 Nd and 155 Gd for Sm isotopes) were measured with 10 12 Ω amplifiers.Sample solutions were introduced into the plasma source via a Teledyne Cetac Aridus3 desolvation system, and the plasma source was equipped with a standard Ni sampler and h-type skimmer cones. 143Nd/ 144 Nd and 87 Sr/ 86 Sr isotope ratio determinations are based on 100 mass scans obtained in static ion-collection mode.Measurements of blank solutions (0.5 M HNO 3 ; 40 scans) preceded and followed those of sample solutions and references materials.
Data reduction was performed off-line using a Python script employing the 'double' isotope dilution equation of Boelrijk (1968).The average value of the two blank solution analyses was used for correction of acid and gas blank contributions and electronic background.For Sr isotope analyses, the blank-and 87 Rb-corrected Sr isotope ratios were normalized to 86 Sr/ 88 Sr = 0.1194 whereas for Nd isotope analyses, the blank-and 144 Sm, 150 Sm-corrected Nd isotope ratio were normalized to 146 Nd/ 144 Nd = 0.7219 (DePaolo, 1981) using an exponential mass fractionation law.Concentrations of Rb and Sm were determined using an ID-SF-ICPMS technique as described in Willbold and Jochum (2005).
143 Nd/ 144 Nd ratios are also expressed here in epsilon (ε) notation, i.
e. ε Nd (0) and ε Nd (T), a measure of the difference in 143 Nd/ 144 Nd from the 'chondritic uniform reservoir' (CHUR), and were calculated using the present-day CHUR values ( 143 Nd/ 144 Nd = 0.512630; 147 Sm/ 144 Nd = 0.1960) of Bouvier et al. (2008).For a rock of age T, ε Nd (T) is the initial value of ε Nd , i.e. the value at the time of formation of the rock, whereas ε Nd (0) is the measured value.Although the depositional ages of the Ediacaran mudstone samples have some uncertainty, this does not change the geological interpretation of the data appreciably.
The 147 Sm/ 144 Nd isotopic ratio is expressed as an enrichment factor (f Sm/Nd ) relative to CHUR and was calculated following DePaolo and Wasserburg (1967): f Sm/Nd = (( 147 Sm/ 144 Nd) sample /( 147 Sm/ 144 Nd) CHUR ) -1.It varies inversely with enrichment of LREE, so it is a sensitive indicator of the REE pattern.The Sm-Nd depleted mantle (DM) crustal residence ages (T DM ) were calculated using the DM evolution model of Nägler and Kramers (1998) and a decay constant (ʎ 147 Sm) of 6.54 × 10 -12 year − 1 (Lugmair and Marti, 1978).The international chronostratigraphic chart (Cohen et al., 2021) was used as a stratigraphic reference for data interpretation.

Results
Based on the analytical data (Tables 1-3), significant variations are observed among the studied samples which define three distinct populations referred to as P1, P2, and P3 in the following.P1 includes samples from the Nyborg and Mortensnes formations.P2 includes the lowermost five samples from the Indreelva Member.All the remaining samples are assigned to P3, extending from approximately 2 m below the stratigraphically lowest occurrence of discoidal Ediacara-type fossils in the Indreelva Member (see Section 1 in Fig. 3), including the Manndrapselva Member and up to the Lower Breidvika member (see Section 2 in Fig. 3).In the following, these three populations will be used when describing and discussing the data.

Mineralogical analysis
Semi-quantitative bulk mineralogy based on XRD is presented in Table 1 and Fig. 5. High amounts of phyllosilicates, mainly illite plus minor amounts of chlorite and biotite, as well as medium amounts of quartz and feldspar are the most prominent minerals in the studied samples.Note that in this study illite data includes muscovite because both 2:1 layer phyllosilicates are difficult to distinguish by semiquantitative XRD analysis on rock powder.Chlorite can vary from Ferich chlorite, known as chamosite, to Mg-rich chlorite, known as clinochlore.Regardless of which phyllosilicate minerals of the chlorite group are present in the samples, they are classified here as chlorite.P1 samples exhibit the highest quartz + feldspar abundances among all samples.The single sample from the Mortensnes Formation contains 3% carbonate (dolomite).

Rare earth element (REE) data
REE abundances are presented in Table 2.The overall REE patterns of P1 samples are similar but the single sample from the Mortensnes Formation shows an overall depletion in REE (Fig. 6a, b).The chondritenormalized patterns of the P1 mudstones exhibit enrichment of the LREE (La N /Sm N : 3.72 to 4.40, with an average value of 4.12; Table 2), relatively flat patterns for the heavy REE (HREE include Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu), with Gd N /Yb N values showing some variation (1.49 to 1.61, with an average value of 1.56).The single sample from the Mortensnes Formation has the lowest La N /Yb N (8.77) compared with La N /Yb N from the other samples of P1 (10.39 and 11.84).The Ce and Eu anomalies range from 0.98 to 1.00 (0.99 on average) and 0.52 to 0.63 (0.56 on average), respectively (Fig. 7a).The PAAS-normalized REE patterns are relatively flat with Nyborg Formation mudstones having higher REE abundances than PAAS while the single sample from the Mortensnes Formation has lower REE abundances than PAAS (Fig. 6b).
The chondrite-normalized patterns of P2 mudstones exhibit enrichment of the LREE (La N /Sm N : 4.73 to 5.23, with an average value of 4.91; Table 2), a depletion of the HREE, with Gd N /Yb N values showing prominent variation (0.94 to 2.78, with an average value of 2.11) (Fig. 6c, d).La N /Yb N of P2 shows overall the highest values among all the studied samples (15.30 to 22.49, with an average value of 19.65; Table 2).The Ce anomalies range from 0.82 to 0.88 (0.85 on average) and hence are lower than those of P1 (Fig. 7a).On the contrary, the Eu anomalies range from 0.64 to 0.85 (0.76 on average) and hence are higher than those of P1.Sample D14-AP-38 shows a pronounced negative Gd anomaly in both the chondrite-and PAAS-normalized REE patterns (Fig. 6c, d).All P2 mudstones exhibit depletion of the HREE with respect to PAAS (Fig. 6d).
The overall REE patterns of P3 mudstones are similar (Fig. 6e-h).For illustration purposes, however, P3 has been split into two subsets.The remaining samples of the Indreelva Member are shown in Fig. 6e, f and those from the Manndrapselva Member and the Lower Breidvika member are shown in Fig. 6g, h.The chondrite-normalized patterns of the P3 mudstones exhibit enrichment of the LREE (La N /Sm N : 3.10 to 6.68, with an average value of 4.74; Table 2), relatively flat patterns for the HREE, with Gd N /Yb N values showing some variation (0.78 to 1.76, with an average value of 1.39).La N /Yb N ranges from 7.45 to 15.73 (10.07 on average).The Ce and Eu anomalies range from 0.96 to 1.01 (0.98 on average) and 0.50 to 0.87 (0.65 on average), respectively.The PAAS-normalized REE patterns are relatively flat (Fig. 6f, h).Sample D14-AP-29 shows a pronounced negative Gd anomaly in both the chondrite-and PAAS-normalized REE patterns (Fig. 6g, h).
Fig. 7b shows Gd N /Yb N plotted against La N /Sm N .Many samples plot near PAAS, one sample of P2 and four samples of P3 show La N /Sm N (>5.0), and four samples of P2 show high Gd N /Yb N (>2.0).

147 Sm-143 Nd isotopic data
Sm-Nd isotopic data are presented in Table 3 and (McLennan and Hemming, 1992;Bock et al., 1994;Lev et al., 1999).A chondritic mantle has an f Sm/Nd of 0.0, LREE-depleted mantle has an f Sm/ Nd of about +0.1 and a typical LREE-enriched upper crustal rock has an f Sm/Nd of about − 0.4 (Taylor and McLennan, 1985).For all samples, regardless of which population they belong, f Sm/Nd values range between − 0.330 and − 0.576 (− 0.476 on average) (Fig. 8b).Rb-87 Sr isotopic data 87 Rb-87 Sr isotopic data are presented in Table 3.With the exception

Table 1
Semi-quantitative powder X-ray diffraction (XRD) bulk mineralogy of mudstones analysed in this study.P* refers to the sample populations identified based on provenance characteristics.The stratigraphic position of the samples is shown in Fig. 3 of the single sample from the Mortensnes Formation, the trace element abundances of Rb and Sr of the mudstones from all three populations show no significant differences.Rb and Sr abundances range from 93 to 220 ppm (with an average value of 158 ppm) and 14 to 132 ppm (with an average value of 53 ppm), respectively.Rb/Sr ratios range from 0.7 to 10.2 (with an average value of 4.1).The single sample from the Mortensnes Formation has the lowest Rb and highest Sr concentrations and hence the lowest Rb/Sr ratio.The Rb-Sr isotopic data yield an errorchron age of ca.430 Ma and an initial 87 Sr/ 87 Sr composition of 0.7352 (Fig. 10).

Mineralogical analysis
Whole-rock XRD analyses reveal that the Ediacaran and Cambrian samples from the Digermulen Peninsula are carbonate-free, with the exception of sample D14-AP-33 from the glaciogenic Mortensnes Formation (belonging to P1) which contains a minor amount of carbonate (dolomite) (Table 1).The carbonate is likely primary since parts of the Mortensnes Formation diamictites comprise clasts and matrix with dolomitic composition (Rice et al., 2011;and own observations).Nonetheless, all data plot on (or close to) the line between the clay and quartz end-members in the ternary diagram for mudstone classification (Fig. 5).All samples have a clay content > 50% and thus are argillaceous mudstones.

Rare earth element (REE) data
Chondrite-and PAAS-normalized REE patterns of the mudstones reveal differences among the samples (Fig. 6).Several interpretations may be put forward to explain the variation in REE patterns (Bock et al., 1994): (1) physical sorting of mineral phases enriched in REE (Cullers et al., 1975;McLennan, 1989;Cullers, 1995), (2) mixing of upper crustal sources with a component with relatively LREE-depleted patterns, e.g.mantle-derived, volcanic material (McLennan et al., 1990), (3) sedimentary modification of the REE during weathering and/or early diagenesis, and (4) modification of the REE during later diagenesis and/ or tectono-metamorphic overprint (Milodowski and Zalasiewicz, 1991;Lev et al., 1999).The first option can play an important role when comparing sediments of different grain sizes, such as mudstones and sandstones and sediments deposited under different depositional environments (e.g., fluvial versus marine).Physical sorting has a greater influence on the chemical signature of sandstone than on mudstone (Cullers, 1995).In this study, sedimentary rocks of similar grain size (mudstone) and of similar depositional environment (marine) were analysed and hence physical sorting of mineral phases can be neglected or may have played only a minor role.The third and fourth options may have played a role but if so, they would have affected all samples and not a selected subset alone.The most likely interpretation to explain the overall variation in the REE patterns (i.e.variations in LREE to HREE abundances) is change in provenance through time, i.e. mixing of upper crustal sources with different REE patterns.This is confirmed by Nd isotopic studies as outlined in Section 5.3.
Compared to all other samples, P2 mudstones show a strong HREE depletion increasing from Gd to Lu (Fig. 6c, d).Zircon is one of the most common accessory minerals in siliciclastic sediments and sedimentary rocks (Pettijohn, 1941) and a main carrier of HREE, showing increasing values from Gd to Lu in normalized REE patterns (Belousova et al., 2002;Rubatto, 2002).Zircon that crystallized under the presence of garnet will show lower HREE concentration than zircon that crystallized under the absence of garnet (Rubatto, 2002;Taylor et al., 2017).Because zircon is stable during weathering and diagenesis (Pettijohn, 1941) postdepositional modification can be excluded and the HREE depletion in P2 mudstones reflects the characteristics of the source area (provenance).The REE depletion observed in the single sample from the Mortensnes  Formation is likely a dilution effect due to carbonate enrichment, as this sample has about 3% dolomite (Table 1).The Ce anomaly in fine-grained sedimentary rocks is commonly used as a fingerprint for the bottom-water redox environment at the time of deposition (e.g., Wilde et al., 1996;Pattan et al., 2005).As Ce has a relatively short mean ocean residence time of only 50-130 years (Alibo and Nozaki, 1999), the Ce anomaly will display a local-regional redox signal (e.g., Shields and Stille, 2001).Changes in bottom-water redox conditions that can influence the Ce anomaly transfer to the sedimentary record may be caused, among others, by sea-level changes (Wilde et al., 1996).Limitations of the application of the Ce anomaly as a palaeoredox proxy may exist in high sedimentation rate areas of continental margin environments (Nath et al., 1997).Increasing dissolved oxygen values in ocean water during the Ediacaran may have provided a first-order control on redox evolution through regulating circulation mechanisms and thus the evolution of the earliest known metazoans (Bowyer et al., 2017).Following the method of Bau and Dulski (1996), P1 mudstones and a few from P3 do not display any Ce anomaly while all P2 mudstones and some from P3 display true negative Ce anomalies (Fig. 7a; Table 2).The true negative Ce anomalies are a clear indication that P2 mudstones and some from P3 were deposited under well-oxygenated seafloor conditions.
The absence of any Ce anomaly in P1 mudstones and few from P3 allows two interpretations.The sediments could have been deposited under anoxic bottom-water conditions.Alternatively, they may have been deposited under oxic bottom-water conditions but in a restricted (near-shore) setting or enclosed bay in a highly evaporitic environment that can preclude Ce anomaly formation (Tostevin et al., 2016a, and references therein).The upper part of the Nyborg Formation was deposited in a tide-influenced setting, indicated by herringbone crossbedding as well as flaser and lenticular bedding (Banks et al., 1971;and own observations).Such an environment is affected by tidal forcing which can lead to extremely high temporal and spatial shifts in oxygen availability in the water column and porewaters (Taillefert et al., 2007).Consequently, although either depositional environment is permissible, the cross-stratified beds in P1 combined with the large negative Ce anomalies dominating P2 and P3 seem to favor oxic bottom conditions throughout the succession.The overall change to a shallow water environment in the upper part of the Nyborg Formation may have been associated with the eustatic sea-level fall which occurred with the onset of the Gaskiers glaciation (Banks et al., 1971;and own observations).
The Eu anomaly in fine-grained sedimentary rocks is a good fingerprint for source rock characterization (e.g., McLennan, 1989;McLennan et al., 1993).Positive or no Eu anomalies are often an indication of REE inputs from plagioclase, because this mineral commonly becomes enriched in Eu during crystallization as Eu 2+ partly substitutes for Ca 2+ (Taylor and McLennan, 1985;McLennan, 1989;Gao and Wedepohl, 1995).Thus, positive or no Eu anomalies in sediments are considered to reflect influx from juvenile undifferentiated arc rocks, characterizing active continental-margin settings (McLennan, 1989;McLennan et al., 1993;Gao and Wedepohl, 1995).Negative Eu anomalies are commonly associated with felsic minerals that precipitate from late-stage magmas, depleted in Eu.Negative Eu anomalies in sediments are considered to reflect influx from either differentiated arc rocks or old upper continental crust or recycled sedimentary rocks, with the latter two characterizing evolved stable cratons (McLennan, 1989;McLennan et al., 1993;Gao and Wedepohl, 1995).The mudstones studied here display variable, significant negative Eu anomalies (Fig. 6).These values suggest that intracrustal differentiation processes such as partial melting or fractional crystallization had affected the source rocks.Thus, differentiated arc rocks, old upper continental crust and/or recycled sedimentary rocks supplied most of the detrital material.P2 mudstones seem to have received the least differentiated material compared with the other samples.
Two samples, D14-AP-29 and D14-AP-38, show distinct negative Gd anomalies in chondrite-and PAAS-normalized diagrams (Fig. 6c, d, g,  h).Also, sample D14-AP-26 shows a negative Gd anomaly but it is less pronounced compared to the other two samples.As the overall dataset and normalized REE patterns appear homogenous, we consider the negative Gd anomalies to be true anomalies and not an analytical Fig. 5. Classification of mudstone samples (according to the nomenclature of Lazar et al., 2015) analysed in this study based on powder XRD bulk mineralogy (Table 1).Well-known Palaeozoic mudstones from the USA are shown for comparison (blue squares; Anderson, 2014).artefact.Although not very common, negative Gd anomalies in sedimentary rocks have been described in the literature.Bhandari et al. (1996) studied samples from the Cretaceous-Tertiary boundary clay layer at Anjar in the Kutch district of western India.Two of their samples from the boundary interval show negative Gd anomalies in both chondrite-and shale-normalized REE diagrams.Also, mangrove sedimentrich in dissolved organic carbon (Orem et al., 1986;Dittmar et al., 2006) sometimes shows a negative Gd anomaly (Prasad and Ramanathan, 2008).The Gd anomaly may be controlled by ocean water circulation, biogeochemical cycling of nutrients, and some trace metals (Alibo and Nozaki, 1999).As the reason for the negative Gd anomaly is not clearly understood, it is not further discussed here.
In a La N /Sm N versus Gd N /Yb N diagram (Fig. 7b), the influence of the enrichment or depletion of specific heavy minerals on the REE abundances in the sediment can be studied (Bock et al., 1994).Many samples plot near PAAS have an average shale composition.Five samples have elevated La N /Sm N , likely related to a depletion in apatite.Four samples from the group P2 have Gd N /Yb N >2.0; in general, the Gd N /Yb N of sedimentary rocks is rather constant and ranges between 1.0 and 2.0 (McLennan et al., 1993).Adding even a small amount of monazite can result in a significant increase of Gd N /Yb N in the sediment (McLennan et al., 1993).Monazite is a phosphate mineral with high LREE content (Deer et al., 1992) and commonly between 0.5 and 4 % Gd 2 O 3 (Lee and Bastron, 1967;Milodowski and Zalasiewicz, 1991;Catlos and Miller, 2017;Lazareva et al., 2018).In case of monazite enrichment, one would expect elevated LREE (especially Ce) and Gd in the absolute element abundances of these samples which is not the case (Table 2).Because all groups show similar PAAS-normalized LREE (Fig. 6) but only P2 shows depletion in HREE, which can be most conveniently explained by a missing HREE component, i.e. zircon.This leads to high Gd N /Yb N in four out of five P2 samples (see inset in Fig. 7b) and reflects source area characteristics.

147 Sm− 143 Nd isotopic data
For all samples, the negative ε Nd (T) values (− 6.00 to − 21.04; Table 3) calculated for the mudstones indicate a dominant input of terrigenous detritus with an old continental crust affinity (Fig. 8).P3 samples seem to have some contributions from minor juvenile sources,  et al., 1994).The inset shows the effects of heavy mineral sorting for apatite, monazite, and zircon, with the minus and plus symbols indicating loss and addition, respectively.Normalization was carried out against CI chondrite (Pourmand et al., 2012).Post-Archaean Australian Shale (PAAS, values from Pourmand et al., 2012) shown for reference.Samples around the PAAS composition appear "normal" in regard of their REE composition.Samples away from the PAAS composition suggests that these samples have been influenced by a REE redistributing process.(1992).Vertical variation of f Sm/Nd can be explained by depletion or enrichment of LREE (Bock et al., 1994;Lev et al., 1999).as these samples show the least negative ε Nd (T) values.The vertical variation in f Sm/Nd within one population of samples (Fig. 8) can be explained by LREE enrichment and depletion, respectively, during crustal formation of the source lithologies (Bock et al., 1994;Lev et al., 1999).Calculated Sm-Nd depleted mantle (DM) crustal residence times (T DM ) of siliciclastic sedimentary rocks can be considered as a weighted average of the age and provenance of the eroded rock units in the source area (e.g., O'Nions et al., 1983;Goldstein et al., 1984;McLennan and Hemming, 1992).For the studied mudstones of the Digermulen Peninsula, T DM crustal residence ages confirm the input of detrital material from old continental crust (Fig. 9).Present-day distribution of basement rocks is only an approximation of that during the time of deposition of the studied sediments as erosion took place since late Ediacaran times.Nevertheless, assuming that the majority of the detrital material was derived from the Fennoscandian Shield of the East European Craton (Fig. 11), possible source areas can be deciphered based on their geochronological ages and Nd isotopic composition.
P1 samples, including the Nyborg and Mortensnes formations, yield mid-Palaeoproterozoic T DM crustal residence ages (Fig. 9).These, together with the negative ε Nd (T) values, reflect the isotopic signatures of felsic igneous rocks from the Svecofennian province (Figs. 9,11).Such provenance is supported by palaeocurrent data (Fig. 3), suggesting sediment transport largely from the south (present-day coordinates).Besides, a possible sediment supply from the Svecofennian province is also supported by the fact that detrital zircon U-Pb ages from a sandstone sample of the Nyborg Formation from the adjacent Varanger Peninsula show mainly mid-Palaeoproterozoic ages (Roberts and Siedlecka, 2012).
P2 samples, the oldest part of the Indreelva Member, yield notably older crustal residence ages (T DM ) (Fig. 9) that reflect sediment supply from mainly early Palaeoproterozoic and locally Neoarchean rock units.The sediment source must be sought in the Karelian province (Figs. 9,11).Thus, in a very broad picture, P1 sediment, deposited during a glacial interval, has a more distal source compared to P2 sediment, deposited in post-glacial time, with a more proximal source.
The remaining sedimentary succession (P3), starting from about 2 m below the stratigraphically lowest occurrence of discoidal Ediacara-type fossils to the Lower Breidvika member, yields late Palaeoproterozoic to early Mesoproterozoic crustal residence ages (T DM ) (Fig. 9).P3 sediment received detrital material from upper Palaeoproterozoic to lower Mesoproterozoic upper crustal rocks of the Svecofennian province, probably including a lot of recycled clastic sediment, and in addition likely with mixing of juvenile detrital material as the P3 samples have the least negative ε Nd (T) values (Fig. 8).Such juvenile detrital material may come from young (volcanic) rocks.This is supported by the fact that detrital zircon U-Pb ages from uppermost Ediacaran and lower Cambrian sandstones on the mainland of northern Norway show evidence for mixed sediment influx from old upper crustal sources from the Svecofennian province and juvenile detrital material (ca.550-570 Ma), the latter being interpreted to be derived from magmatic rocks of the Timanides (Andresen et al., 2014;Zhang et al., 2015).Overall,   (Vermeesch, 2018).Overdispersion is attributed to the geological scatter in the ages or in the non-radiogenic isotope composition.
significant shifts in Sm-Nd isotope values obtained from mudstones of the Digermulen Peninsula are in agreement with published detrital zircon dates from sandstones of northern Norway (Roberts and Siedlecka, 2012;Andresen et al., 2014;Zhang et al., 2015), reveal prominent changes in sediment sources (Fig. 12), probably related to large-scale palaeotectonic reorganisation along the Iapetus-facing margin of Baltica during late Neoproterozoic and early Cambrian times.

87 Rb-87 Sr isotopic data
The combined Rb-Sr isotopic data of all samples yield an errorchron age of about 430 Ma (Fig. 10).This is consistent with resetting of the Rb-Sr whole-rock isotope systems of the mudstones during a known tectono-metamorphic event in the Gaissa Nappe Complex of Finnmark (Dallmeyer et al., 1989), likely representing the peak of metamorphism and cleavage generation on the Digermulen Peninsula.Nonetheless, interpreting Rb-Sr isotope data in sedimentary rocks is highly speculative because of the mobility of Rb and Sr even during diagenetic processes.Previous geochronological data ( 40 Ar/ 39 Ar, K-Ar, Rb-Sr) of whole-rock from the Gaissa Nappe Complex are not easily gauged and have remained controversial (Sundvoll and Roberts, 2003).In Norway, late Silurian-Early Devonian metamorphic ages are related to the continental collision between Baltica and Laurentia and the closure of the Iapetus Ocean, the so-called Scandian phase of the Caledonian Orogeny (Gee, 1975;Roberts, 2003).In Finnmark, about 431-428 Ma-old metamorphic ages are related to main phase of the Scandian Orogeny before gravitationally driven collapse of the orogen (Kirkland et al., 2006).The Rb-Sr age obtained in this study overlaps in error with a  2001) and insert map based on Gorbatschev and Bogdanova (1993).TIB -Transscandinavian Igneous Belt.calcite U-Pb age of 418 ± 23 Ma from deformed cone-in-cone calcite within the 2nd cycle of the Manndrapselva Member on the Digermulen Peninsula (Meinhold et al., 2020).It also overlaps in error with a Rb-Sr age of 430 ± 11 Ma from the<0.1 μm grain-size fraction from the Indreelva Member of the parautochthonous units on the Varanger Peninsula (Gorokhov et al., 2001).However, the Rb-Sr ages are about 30-40 Ma older than K-Ar ages of 403.9 ± 4.2 and 391.5 ± 4.0 Ma from the < 2 and < 0.2 μm grain-size fractions, respectively, of the upper Indreelva Member on the Digermulen Peninsula; these ages have been interpreted to present a thermal overprint under low epizonal conditions along a localised shear zone during the late stage of the Scandian Orogeny (Meinhold et al., 2019b).In summary, at least two tectonometamorphic phases affected the sedimentary rocks from the Digermulen Peninsula in Finnmark: the main phase and the late phase of the Scandian Orogeny.

Implications for Ediacaran biological evolution
Macroscopic metazoans could only have evolved once permissive levels of seawater oxygen had been established, though it is unclear by what time such conditions were reached and how stable they were.However, it has to be factored in that early metazoanssome found among the Ediacara biotamay have had a high tolerance to low oxygen concentrations under which they evolved, as do some modern animals (for instance, sponges; Mills et al., 2014).This notwithstanding, it is generally assumed that the abundance and increase in complexity of the Ediacara biota is related to increased oxygen levels (Sperling et al., 2016;Evans et al., 2018).In most models, an increase in seawater oxygenation significantly precedes the appearance of animals in the fossil record.However, Shi et al. (2022) found evidence that oceanic oxygenation coincided with the Shuram excursionthe largest known negative carbon isotopic excursion in Earth's history that occurred between 574.0 ± 4.7 Ma and 567.3 ± 3.0 Ma (Rooney et al., 2020) approximating the appearance of the earliest Ediacara biota.Recent studies suggest modest and fluctuating oxygen levels during the Ediacaran Period (e.g., Tostevin and Mills 2020), which may have controlled the distribution of animals (e.g., Li et al., 2015;Cui et al., 2016).
Distribution of the Ediacaran macrofossils in successions in China, Namibia and Siberia is largely controlled by the redox state of the water column (e.g., McFadden et al., 2008;Wood et al., 2015;Bowyer et al., 2017).Large and complex organisms, especially taxa of the earliest mobile metazoans may have had metabolic demands for dissolved oxygen concentration higher than during most of the Precambrian (e.g., Sperling et al., 2013;Tostevin et al., 2016b).Along these lines, Evans et al. (2018), based on data from South China, concluded that the appearance of morphologically complex Ediacara-type fossils of the White Sea association (approximately 560-551 Ma) coincided with increased seawater oxygen levels.They also suggested that the lower diversity in the succeeding Nama assemblage (approximately 551-540 Ma) coincided with, and was caused by, a decrease in oxygen levels.Zhang et al. (2019) found evidence for an earlier appearance in the rise of seawater oxygen but similarly found evidence for increased sea-floor anoxia at the terminal part of the Ediacaran.The terminal Ediacaran decline in biotic diversity on the East European Platform has been called the 'Kotlinian Crisis' (Kolesnikov et al., 2015).Grazhdankin (2014) attributed this decline to the rise of mobile bilaterians and no evidence was detected for environmental changes and bottom water redox conditions remained unchanged through the biotic change recorded on the East European Platform (see Maslov et al., 2019).
We compared the sedimentological evidence for the change in oxygenation state of the bottom water with the fossil record observed in the older strata of the Stáhpogieddi Formation (mainly lower part of the Indreelva Member) to see whether environmental conditions may have been important for early macroscopic life on Baltica.Proxy results from P2 and some P3 mudstones suggest that the earliest appearance of macroscopic fossils in the basal part of the Indreelva Member (e.g., palaeopascichnids and probable metazoan Ediacara-type fossils) in Finnmark/Scandinavia (Högström et al., 2013;Jensen et al., 2018b) was associated with sediment deposition in a well-oxygenated environment.This interval is probably younger than approximately 560 Ma (Jensen et al., 2018b), so temporarily belongs to the White Sea association.The differences in the fossil assemblages between the two principal fossiliferous levels in the Indreelva Member can be attributed to facies.The Indreelva Member has been classified as an example of Fermeuse-style preservation which characterizes deeper water settings in which microbial mats were not normally formed reducing the preservation of surficial fossils, with a predominance of Aspidella discs formed in the sediment (Narbonne, 2005).Specimens of Aspidella in the lower Indreelva Member (Fig. 4a) are associated with the appearance of thin sandstone beds.These Aspidella often show a depth of several centimetres, are consistent with an organism's partial emplacement in the sediment.The more diverse assemblage of fossils typical of the White Sea assemblage higher in the Indreelva Member (Fig. 4b) is attributed to a facies more conducive to their preservation.The results from redox-sensitive elements presented in this paper allow evaluating the role of the oxygenation state of the bottom water on the distribution of macroscopic fossils.The onset of oxic depositional conditions observed in the P3 sediments coincides with the occurrence of the Ediacara-type biota.Our data further suggest bottom water oxygenation for the latest Ediacaran of the Manndrapselva Member and the top of the Indreelva Member.This part of the succession yields palaeopascichnids (Fig. 4c) but metazoan Ediacara-type fossils have not been found.Trace fossils are prominent in the Ediacaran middle part of the Manndrapselva Member (McIlroy and Brasier, 2017;Jensen et al., 2018b), typical of the Nama-type time-equivalent associations elsewhere (e.g., Darroch et al., 2021).Our results are similar to those of earlier studies from the East European Platform (Johnston et al., 2012;Maslov et al., 2019) in suggesting oxic seafloor conditions during the Ediacaran, suggesting that other factors likely were more important for the pattern of biotic changes, such as nutrients (e.g., Pehr et al., 2018), competition (Grazhdankin, 2004), or inevitable outcomes of evolutionary processes (cf.Budd and Mann, 2020).On the other hand, the shift in sediment provenance, i.e. change in lithological composition of the source area(s), between P1, P2 and P3 is probably not responsible for the diversity and composition of fossil assemblages.For example, palaeopascichnids occur in P2 and P3 sediments.However, the shift in sediment provenance may have influenced the sedimentation rate or the depositional setting, which would have an impact on the preservation of fossils.

Conclusions
The analyses of the Nd and Sr isotopic and REE composition of the fossil-bearing Ediacaran and lowermost Cambrian succession of the Digermulen Peninsula, Arctic Norway, provide first constraints on the provenance and post-depositional alteration of these strata.Our study shows that: 1) All studied samples are argillaceous mudstones.They are carbonate-free, with the exception of small amounts in a single sample from the glaciogenic Mortensnes Formation.
2) Mudstones from the Nyborg and Mortensnes formations do not display any Ce anomaly.They yield mid-Palaeoproterozoic T DM crustal residence ages, and together with the negative ε Nd (T) values reflect the isotopic signatures of felsic igneous from the Svecofennian province.
3) The lowermost five samples from the Indreelva Member display true negative Ce anomalies, and thus sediments were deposited under well-oxygenated seafloor conditions.Their crustal residence ages (T DM ) reflect sediment supply from mainly early Paleoproterozoic and locally Neoarchean rock units, as present in the Karelian province.
4) Samples starting about 2 m below the stratigraphically lowest occurrence of discoidal Ediacara-type fossils to the Lower Breidvika member show varying Ce anomalies.Their negative ε Nd (T) and T DM values suggest sediment supply from upper Palaeoproterozoic to lower Mesoproterozoic upper crustal rocks of the Svecofennian province, and in addition likely some juvenile (late Neoproterozoic volcanic) source.
5) Shifts in Sm-Nd isotope values from the glacial to the post-glacial succession and within the post-glacial succession, approximately 2 m below the first appearance of macroscopic Ediacara-type fossils of probable metazoans, are related to changes in the sediment source and probably to large-scale palaeotectonic reorganisation along the Iapetusfacing margin of Baltica.
6) The combined Rb-Sr isotopic data of all samples yield an errorchron age of about 430 Ma reflecting the resetting of the Rb-Sr wholerock isotope systems of the mudstones during a Scandian tectonometamorphic event in the Gaissa Nappe Complex of Finnmark.
7) The earliest appearance of macroscopic fossils in Finnmark/ Scandinavia was associated with sediment deposition in a welloxygenated environment.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 2 .
Fig. 2. Geology of the southeast portion of the Digermulen Peninsula, based on Siedlecka et al. (2006), showing localities of the two sections where samples were taken for analyses discussed in this study.The inset (upper left) indicates the location of the geological map on the Digermulen Peninsula.Section 1 comprises the older part of the succession, including the uppermost part of the Nyborg Formation, Mortensnes Formation, and the lower part of the Indreelva Member.No mudstone samples were taken from the Lillevannet Member due to its coarse lithology.The geographic coordinates of the base and top of Section 1 are as follows: N 70 • 33 ′ 51 ′ ′ and E 028 • 06 ′ 17 ′ ′ for the base, and N 70 • 34 ′ 18 ′ ′ and E 028 • 07 ′ 47 ′′ for the top.Section 2 comprises the uppermost part of the Indreelva Member, Manndrapselva Member, and the lower part of the Lower Breidvika member.The geographic coordinates of the base and top of Section 2 are as follows: N 70 • 34 ′ 43 ′ ′ and E 028 • 08 ′ 60 ′ ′ for the base, and N 70 • 34 ′ 58 ′ ′ and E 028 • 08 ′ 31 ′ ′ for the top.

Fig. 4 .
Fig. 4. Examples of late Ediacaran fossils from the Digermulen Peninsula, Arctic Norway.(a) Aspidella-type discoidal fossils from the basal part of the Indreelva Member in a coastal outcrop north of Árasulluokta.See section 1 in Fig. 3 for stratigraphic level.Scale bar is 50 mm (b) Multi-vaned frondose fossil (cf.Swartpuntia) from the approximate mid-portion of the Indreelva Member in a coastal outcrop north of Manndrapselva.Scale bar is 20 mm.(c).The non-metazoan fossils Palaeopascichnus and Harlaniella from the 2nd cycle of the Manndrapselva Member.See section 2 in Fig. 3 for stratigraphic level.Scale bar is 10 mm.
illustrated in Figs. 8, 9. Significant variations are observed among the three populations.P1 has 143 Nd/ 144 Nd initial ratios ranging from 0.511250 to 0.511305 (with an average value of 0.511285), ε Nd (0) values ranging from − 18.09 to − 18.65, ε Nd (T) values ranging from − 11.08 to − 12.40, and 147 Sm-143 Nd depleted mantle (DM) crustal residence ages (T DM ) of 1830 to 2045 Ma.P2 samples have a significantly different isotopic composition.They have 143 Nd/ 144 Nd initial ratios ranging from 0.510834 to 0.510868 (with an average value of 0.510850), the most negative ε Nd (0) values, ranging from − 25.64 to − 28.30, ε Nd (T) values ranging from − 20.37 to − 21.04, and the oldest T DM ages, ranging from 2360 to 2840 Ma.P3 has 143 Nd/ 144 Nd initial ratios ranging from 0.511804 to 0.511991 (with an average value of 0.511902), ε Nd (0) values ranging from − 12.46 to − 16.11, ε Nd (T) values ranging from − 6.00 to − 9.01, and T DM ages of 1405 to 1960 Ma.Fig. 8a illustrates the separation of the stratigraphic succession into three units based on the Sm-Nd isotopic signatures.Fig. 8b combines REE abundances and Nd isotopic composition, with f Sm/Nd being an expression of the slope of the REE pattern between Nd and Sm

Fig. 6 .
Fig. 6.Normalized REE patterns of mudstones analysed in this study.Normalization was carried out against CI chondrite and Post-Archaean Australian Shale (PAAS) using the revised CI chondrite and PAAS values of Pourmand et al. (2012).(a-d) Mudstones from Ediacaran glacial-interglacial cycle.(e-h) Mudstones across the Ediacaran-Cambrian boundary.

Fig. 7 .
Fig. 7. (a) Plot of Pr/Pr* versus Ce/Ce* (after Bau and Dulski, 1996) to reveal genuine Ce anomalies.The distinct fields are defined as follows: 1 -neither Ce nor La anomaly; 2a -positive La, no Ce anomaly; 2b -negative La, no Ce anomaly; 3a -positive Ce and negative La anomaly; 3b -negative Ce and positive La anomaly; 4a -negative Ce and negative La anomaly; 4b -positive Ce and positive La anomaly.(b) Plot of La N /Sm N versus Gd N /Yb N (after Bocket al., 1994).The inset shows the effects of heavy mineral sorting for apatite, monazite, and zircon, with the minus and plus symbols indicating loss and addition, respectively.Normalization was carried out against CI chondrite(Pourmand et al., 2012).Post-Archaean Australian Shale (PAAS, values fromPourmand et al., 2012) shown for reference.Samples around the PAAS composition appear "normal" in regard of their REE composition.Samples away from the PAAS composition suggests that these samples have been influenced by a REE redistributing process.

Fig. 8 .
Fig. 8. (a) Plot of εNd(T) versus stratigraphic ages of the samples.The samples (without the prefix 'D14-') are arranged from old (base) to young (top).(b) Plot of εNd(T) versus f Sm/Nd , with reference fields after McLennan and Hemming

Fig. 10 .
Fig.10.Rb-Sr whole-rock isochron diagram for mudstones analysed in this study, produced using IsoplotR(Vermeesch, 2018).Overdispersion is attributed to the geological scatter in the ages or in the non-radiogenic isotope composition.

Fig. 12 .
Fig. 12. Summary figure showing the stratigraphy of the studied mudstones from the Digermulen Peninsula, the first occurrences of some prominent macrofossils, the obtained εNd(T) values, and the suggested sediment sources.The three populations, labelled P1, P2, and P3, are explained in Sections 4 and 5.

Table 2
Rare earth element (REE) abundances in parts per million (ppm) of mudstones analysed in this study.P* refers to the sample populations identified based on provenance characteristics.The stratigraphic position of the samples is shown in Fig.3.Sm and Nd values were obtained by isotope dilution thermal ionization mass spectrometry (ID-TIMS).All other REE data were obtained by ICP-MS.

Table 3
Rb-Sr and Sm-Nd whole-rock isotopic data of mudstones analysed in this study.P* refers to the sample populations identified based on provenance characteristics.The stratigraphic position of the samples is shown in Fig.3.Column 'Depositional age (Ma)' indicates the approximate stratigraphic age used to calculate ε Nd (T) values.Rb and Sr values were obtained by isotope dilution thermal ionization mass spectrometry, Sr using a Neptun multi-collector inductively coupled plasma mass spectrometer and Rb using a ThermoFisher Scientific Element2 single-collector inductively coupled plasma mass spectrometer.The 'external' relative standard deviation (1RSD) of Rb/Sr (using several sessions) is 0.4% (relative), the 1RSE of individual analysis is about 0.2% (relative).Analytical details about the isotope measurements are given in Section 3.