Detrital zircon records variable expressions of Paleoproterozoic (2.2 - 2.4 Ga) glaciation between Sclavia and Superia supercratons

ABSTRACT Evidence of glaciation is recognized in Earth’s history from 2.9 Ga to the present, providing important information regarding the climatic, sedimentary, and tectonic evolution of our planet. Glaciers are highly effective agents of erosion that generate and transport large amounts of sediments for 100ʹs of kilometres. Therefore, detrital zircon geochronology of glaciogenic sedimentary rocks can elucidate the provenance and evolution of the sedimentary record through time. In this study, we investigate the Paleoproterozoic Glaciation Event which is hypothesized to be the oldest global glaciation in Earth’s history (2.4 to 2.2 Ga). We identify distinct sedimentary provenance patterns between the Archaean supercratons Sclavia and Superia, which, comprised most of the extant Archean crust. The detrital zircon of Sclavia exhibit increasing dissimilarity through time whereas the detrital zircon ages of Superia show consistent age peaks through time. Coupled with the absence of glacial deposits and predominance of marine sedimentary successions within Sclavia, the detrital zircon provenance likely indicates limited subaerial exposure of continental crust in Sclavia with glaciers narrow mountain belts producing no diamictites. This model resolves previous issues related to hypotheses proposing high-obliquity low latitude glaciation of Superia as the only explanation available.


Introduction
Glaciers are effective agents of erosion (Winter et al., 2012). Glacial erosion is defined by the removal and transport of bedrock sediments by abrasion and quarrying by ice along with erosion by glacial runoff (Hambrey et al. 2004). Several studies suggest that glacial erosion is commonly more rapid than fluvial erosional mechanisms (Guymon 1974;Gurnell et al. 1996;Hallet et al. 1996). Multiple large-scale glaciations are recorded throughout Earth's history and their occurrence strongly influences the hydrological cycle, fluvial systems, and the sedimentary record. Two of the largest glacial episodes are theorized during the Paleoproterozoic and Neoproterozoic (Cryogenian Glaciation) eras (Kirschvink et al. 2000;Hoffman and Schrag 2002;Young 2013) during which it is hypothesized that the glaciation extended to low latitudes. These intervals are known as the episodes of Snowball Earth.
The Paleoproterozoic Glaciation, widely known as Huronian Glaciation, is recorded during the earliest Paleoproterozoic. The Archean-Paleoproterozoic transition (2.5 to 2.2 Ga) represents an intriguing period that records the increase of atmospheric oxygen (Canfield 2005;Barley et al., 2005), continental freeboard (Bindeman et al., 2018;Liebmann et al. 2021), diversity of metamorphic conditions (Holder et al. 2019;Brown et al. 2021), evolution of oxygenic photosynthesis (Ward et al. 2016), and tectonic magmatic lull (Condie et al. 2009(Condie et al. , 2022Spencer et al. 2018). The coincidence of these rapid shifts may indicate a coupling between geodynamic processes and supracrustal/atmospheric processes (Spencer et al. 2018). The Snowball Earth model implies a major interval of glacial activity during the Paleoproterozoic (2.4 Ga) (Kirschvink et al. 2000). This theory is widely accepted during this period, considering the distribution of glacial deposits in different cratons (Liu et al. 2021). Also, Paleoproterozoic glaciers is related with the presence of tropical rather than polar latitude evidence, such as red beds, evaporites and carbonate rocks (Kirschvink et al. 2000). Also, paleomagnetic results indicate exposures of the Paleoproterozoic glaciation occurring within ≈11° of the equator (Evans et al. 1997). Paradoxically, Young (2019) consider that the Huronian Supergroup, the typearea for the Paleoproterozoic Glaciation, provides scarce geological evidence to support the Snowball Earth hypothesis. As an example, the occurrence of varve-like deposits in the Gowganda Formation indicates a seasonal character, which is inconsistent with extremely cold conditions required by the snowball Earth scenario (Bekker et al. 1999;Kump et al. 2013). Although many efforts were made to better constrain the global glaciations on Earth, the magnitude and mechanisms of these events are still widely debated.
The Paleoproterozoic Glaciation is posited to represent the world's first widespread glaciation on Earth (Kirschvink et al. 2000;Hoffman 2013;Tang and Chen 2013;Young 2019). In the period from 2.45 Ga to 2.22 Ga, three glaciation cycles are recorded in the Huronian Supergroup, in the eastern portion of Canada (Kopp et al. 2005;Rasmussen et al. 2013). These are represented by the diamictite of the Ramsay Lake, Bruce and Gowganda formations (Kopp et al. 2005). Other units are described in North America, such as Paleoproterozoic glaciogenic deposits in the Marquette Range Supergroup (Chocolay Group) in Michigan and Wisconsin (Bekker et al. 2006) and the Snowy Pass Supergroup in Wyoming. The oldest Huronian glacial unit (Ramsay Lake Formation) is commonly correlated to the Polisarka Formation, located in Immandra-Varzuga Greenstone Belt in the Kola Peninsula, Russia . Furthermore, the Polisarka Formation corresponds to the Neverskrukk Formation in the Pechenga Greenstone Belt, the Urkkavaara Formation (North Karelia Schist Belt) in Finland, the Sariolian Group (Ojakangas et al. 2001), and the Sompujärvi Formation (Perapähja Schist Belt) (Melezhik 2006). The glaciation event in the Kola Peninsula is constrained between 2.4 Ga and 2.2 Ga (Amelin et al. 1995;Hanski et al., 2001). Moreover, the Huronian Supergroup glacial units are interpreted as chrono-correlated to the Duistchland and Boshoek formations in Traansval Supergroup and the Makganyene Formation in Griqualand West Basin, both located in the Kapvaal Craton in South Africa (Rasmussen et al. 2013) (Figure 1). São Francisco, Dharwar, Zimbabwe, Slave, and Yilgarn cratons paleogeographically associated with the supercraton Sclavia also present sedimentary successions deposited during the Paleoproterozoic glaciation timeframe (Supplementary Material - Fig. A1). However, none of these units record any glacial deposits (Liu et al. 2021).
supercraton model suggests two spatially disconnected main landmasses during this period, known as Superia and Sclavia (Figure 2b). Liu et al. (2021) suggested the supercraton solution as the preferred one arguing that the supercontinent solution is weakly supported by the poles of individual ages.
Importantly, preserved glacial deposits are present on all the known cratons of Superia (Superior, Kaapvaal, Kola/Karelia, Wyoming, Hearne, Pilbara) and completely absent on those in Sclavia (Bleeker 2003;Pehrsson et al. 2013;Gumsley et al. 2017). The paleomagnetic analysis also indicates that the cratons recording glaciogenic sedimentary successions have been situated in lowlatitude regions (Evans et al. 1997;Liu et al. 2021). The occurrence of glaciation is primarily located nearer the poles and at high elevations. Therefore, glaciation in the equatorial region is unusual as the equatorial region today is characterized by higher temperatures.
The presence of low-latitude glaciation may be explained either as an episode of global glaciation (Evans et al. 1997;Kirschvink et al. 2000) or a period of highobliquity (Williams et al. 1998;Young 2014). The highobliquity hypothesis suggests a high-obliquity degree (49° to 54°) between the magnetic field and the spin axis. proposing that equatorial regions were colder during the Proterozoic (Williams et al. 1998). The high obliquity theory is posited to have occurred due to a massive bolide impact at ~4.5 Ga in the Earth, that favoured mechanism for lunar origin (Williams 2008). Therefore, high obliquity may have prevailed during the Precambrian with low paleolatitude of glaciations in the Paleoproterozoic and Neoproterozoic (Williams 2008). However, a dynamical mechanism is required to bring the obliquity down to its present value of 23.58°, and this mechanism is still unknown (Donnadieu et al. 2002).
To test the supercraton versus supercontinent hypotheses proposed by Liu et al. (2021), this paper presents compiled detrital zircon ages from sedimentary successions that are pre-glaciation (>2.5 Ga), syn-glaciation (2.5-2.2 Ga) and post-glaciation (>2.2) from nine cratons to better understand changes in detrital zircon provenance, transport of sediment from source to sink, and configuration of the cratons during the Paleoproterozoic glaciation. Given that glaciation significantly impacts erosional processes, the investigation of detrital zircon ages can provide important information regarding changes in sedimentary dynamics and preservation through time. Here, we identify two distinct patterns of zircon age frequency between Sclavia and Superia implying that glacial sedimentary processes provide a diagnostic signature different than non-glacial processes. These findings suggest the hypotheses that Paleoproterozoic glaciation was not recorded in Sclavia supercraton because glaciation is restricted to Superia supercraton or due to submerged landmasses in Sclavia cratons.

Data Compilation and Results
Here we present a compilation of detrital zircon U-Pb age data from nine cratons (n = 7683) (Table A1:  et al. 2021). We visualize the detrital zircon data using kernel density estimation (KDE) and multi-dimensional scaling analysis (MDS) (Vermeesch 2013) ( Fig. 3 and 4). KDE provides an indication of zircon frequency using a fixed bandwidth. MDS visualizes the similarity of zircon age population frequencies between samples by using a dissimilarity matrix with the D value from the Kolmogorov-Smirnov (K-S) test of zircon ages (excluding uncertainties). The matrix of pairwise (Euclidian) distances shows the similarity or the 'distance' between each sample (Vermeesch 2013). Analyses from individual units were combined and classified as a single metric (e.g pre-glacial units, Syn-glacial units and Post-glacial units in Superia or Sclavia). For each case study, we also use synthetic age end members to evaluate dominant age populations in given samples (Spencer and Kirkland 2016). For the calculation of end members, we defined the most representative ages in each metric and calculated 1000 synthetic ages (e.g. 2200 ± 50 Ma). The synthetic unimodal age spectra support the visual comparison of the MDS maps. This provides a vector on the map that can be associated with the increase or decrease in contribution of a given age component (Spencer and Kirkland 2016). Dm1 represents the Distance/Disparity while Dm2 characterizes dissimilarity (Vermeesch 2013). The points of the dissimilarity matrices are contoured using bivariate kernel density estimation (Botev et al. 2010). KDE and MDS plots are shown to compare the distribution of detrital zircon contribution during preglaciation (>2.5 Ga), syn-glaciation (2.4-2.2 Ga) and post-glaciation (<2.2 Ga) period in Superia and Sclavia, respectively ( Fig. 3 and 4). The pre-glacial successions of Sclavia exhibit age dominant peaks at 2.7 Ga and 2.9 Ga, with minor peaks in 2.5 Ga and 3.2 Ga. This signature is similar to the detrital zircon contribution in Superia pre-glacial units. Syn-glacial sediments from Sclavia and Superia show the same detrital contribution (2.7, 2.5 and 2.4 Ga), but their proportions are reverse, since in Sclavia the main peak is associated with the 2.5 Ga grains and 2.7 Ga dominant in Superia. Postglacial units in Superia have as its main age component 2.7 Ga grains while in Sclavia the main contribution corresponds to a 2.1 Ga dominant peak. Therefore, a persistent main peak of 2.7 Ga is showed in Superia, whereas to Sclavia the youngest component is dominant through time. The Superia MDS contours show significant overlap between the pre, syn and post-    glacial data as evidenced by the 50 th and 90 th percentile contour lines, while in Sclavia data contours are nonoverlapping (Figure 3). Individual KDE plots for São Francisco, Wyoming, Fennoscandian and Kaapvaal cratons are also presented in Fig. A2 (Supplementary Material).

Discussion
Detrital zircon populations from the Superia supercraton reinforce the presence of glacial environments and glaciogenic erosional processes. The dominant peak in preglacial sediments (2.7 Ga) in Superia suggests a predominant contribution characterized by a restricted provenance strongly linked to proximal basement rocks. During the deposition of syn-glacial units, the age pattern changes, with a 2.7 Ga main detrital zircon peaks, including another younger peak in 2.5 Ga. Post-glacial multimodal detrital zircon peaks (3.2 Ga, 2.7 Ga, and 2.2 Ga) in Superia indicate the input of multiple sources with a cumulative effect on the detrital zircon ages. A persistent age peak of 2.7 Ga in sedimentary successions of Superia through time can be explained by extensive erosion of pre-existing crust and/or reworking of older sediment successions in glacial environments during ice retreat and deglaciation. The overlapping nature of pre-, syn-, and post-glacial sample density contours in the Superia MDS plot suggests that the transport from source to sink is associated with a variety of detrital zircon populations in a dynamic catchment, common in glaciogenic environments (Craddock et al. 2019). This is in stark contrast to the cratons of Sclavia which show minimal reworking of older sedimentary successions with a clear progression towards the erosion of younger continental material. This pattern can be observed in the provenance histograms with dominant peak pattern, especially in syn and post-glacial units of Sclavia, indicating a limited catchment and localized sources input. This difference in the detrital zircon contribution might indicate a lack of glaciation in Sclavia or a less expressive glacial event.
As discussed above, the lack of glacial sediments in the Sclavia supercraton at more northernly paleolatitudes than the extensive glaciated Superia supercraton can be explained by an episode of highobliquity glaciation. According to paleomagnetic data, the paleolatitude of Sclavia and Superia imply high obliquity for the low latitude Superia glaciation and lack of glaciation at the higher latitudes of Sclavia (Liu et al. 2021). High obliquity theory suggests that the angle between the magnetic field and spin axis has a drastic influence on Earth's climate (Williams 1993;Williams et al. 1998;Young 2013). Williams (1993) proposed that Earth's obliquity was greater than 54° during most of its history, which suggests that the glacial ice from the poles expanded to the tropics causing a Snowball Earth. Alternatively, high-obliquity would cause a climate inversion concerning the latitude, meaning that the low-latitude regions would present the coldest temperatures on Earth, while the poles would be warmer. Although the high obliquity theory elucidates the paradoxical conditions during the low latitude Paleoproterozoic Glaciation, this hypothesis is still under debate, since the mechanism responsible for drastically reducing obliquity to its present value (23°) in a 100 Ma interval is still unknown (Jenkins, 2004).
An alternative hypothesis is that a global glacial event is not recorded on Sclavia due to lack of significant continental freeboard. Assessment of Sclavian sedimentary successions shows a predominance of marine deposits with a dearth of continental sedimentary environments and glacial deposits from 2.5 to 2.0 Ga (Table  A2: Supplementary Material). The limited occurrence of continental deposition could be explained by geological processes, such as burial or weathering removal. Alternatively, it may represent a lack of subaerial deposition and preservation due to submerged Sclavian landmasses. This would imply the presence of isolated mountain chains eroding into the ocean and restricted mountaintop glaciers with a paucity of glaciogenic deposition. According to Bindeman et al. (2018), the area of emerged continental crust increased substantially during this period. Therefore, it is hypothesized that Sclavia and Superia had different proportions of exposed landmasses, implying a contrasting detrital zircon signature between the supercratons. The schematic model presented in Figure 5 provides an explanation for the glaciogenic reworking of older sediments in Superia and the lack thereof in Sclavia. A gap of 2.5 Ga detrital zircon is observed in post-glacial units (>2.2 Ga) in Superia and Sclavia, implying a provenance switch between 2.2 and 2.0 Ga. This could be explained by the reorganization of sedimentary input after a global glaciation.
North China and Amazonian cratons are not included in the supercraton model (Liu et al. 2021). However, Martins et al. (2021) present new evidence indicating a low paleolatitude to the Carajás Basin during 2.75 Ga, which implies that the Amazonian Craton might have been part of the Superia supercraton configuration during the Neoarchean (~2750 Ma) in a position close to the paleoequatorial line. Furthermore, two recent papers described the presence of glaciogenic deposits in the Serra Sul Formation, located in Carajás Province, Amazonian Craton (Araujo et al., 2019;Rossignol et al. 2020) and in Hutuo Group in the North China Craton (Chen et al. 2019). Our detrital zircon compilation shows a persistent peak pattern in 2.7 Ga in the Amazonian (Rossignol et al. 2020;Rossignol et al., 2021;Araújo et al., 2021) and North China cratons (Puetz et al., 2021), implying a sedimentary reworking array comparable to the one observed in the Superia supercraton units (Table A3 and Figure A3: Supplementary Material). More paleomagnetic are required to investigate this hypothesis.

Conclusion
The distinct detrital zircon provenance of Sclavia and Superia during the Paleoproterozoic global glaciation can be explained by the degree of subaerial emergence of these supercratons. In Sclavia, the age of the dominant detrital zircon age peak decreases through time, suggesting restricted sedimentary sources with predominant submerged continental masses and marine sedimentation. In contrast, the supercraton of Superia records the deposition of sediments influenced by substantial glacial erosion, reworking of basement rocks, and deposition of glacially derived sedimentary successions. The lack of Sclavian glaciogenic sedimentary rocks, coupled with paleomagnetic data and compiled U-Pb detrital zircon points toward two hypotheses: (a) the Paleoproterozoic glaciation was regional in nature and not global as presumed by previous research or (b) Sclavia was largely submerged during the Paleoproterozoic glaciation and for this reason it does not record the glacial deposits. Given the requirement of a degree of orbital highobliquity to explain the low-latitude glaciation on Superia and the lack of dynamical models to account for this, we find the latter hypothesis to be most likely. Evaluating the secular change in detrital zircon age populations in the supercratons of Superia and Sclavia allows better visualization of the large role played by glacial erosion and reorganization of sedimentary systems through time. Although several researchers suggest that continental freeboard significantly increased at the  Archean-Proterozoic boundary (Bindeman et al. 2018, Spencer et al. 2019Liebmann et al. 2021), our results demonstrate the plausibility of a diachronous increase of continental freeboard on various cratons/supercratons and therefore more detailed, craton-specific work is needed to constrain the development of continental freeboard during the Archean-Proterozoic boundary.