A very unconventional hydrocarbon play: the Mesoproterozoic Velkerri Formation of Northern Australia.

The ca. 1.5–1.3 Ga Roper Group of the greater McArthur Basin is a component of one of the most extensive Precambrian hydrocarbon-bearing basins preserved in the geological record, recently assessed as containing 429 million barrels of oil and eight trillion cubic feet of gas (in place). It was deposited in an intra-cratonic sea, referred to here as the McArthur-Yanliao Gulf. The Velkerri Formation forms the major deep-water facies of the Roper Group. Trace metal redox proxies from this formation indicate that it was deposited in stratified waters, in which a shallow oxic layer overlay suboxic to anoxic waters. These deep waters became episodically euxinic during periods of high organic carbon export. The Velkerri Formation has organic carbon contents that reach ~10 wt%. Variations in organic carbon isotopes are consistent with organic carbon enrichment being associated with increases in primary productivity and export, rather than flooding surfaces or variations in mineralogy. the Formation an intracontinental setting recent global reconstructions show a broader mid to low latitude gulf, with the Formation coeval of

Petroleum exploration in the ca. 1.5 to 1.3 Ga Roper Group has continued intermittently since the first discovery of 'live' hydrocarbons in the Velkerri Formation in 1985 (Jackson et al., 1986). The two main targets for petroleum exploration in the Roper Group are the Velkerri Formation and the Kyalla Formation Yang et al., 2019;Bodorkos et al., 2020;Yang et al., 2020), both within the Upper Roper Group (the Maiwok Subgroup; Fig. 2). The Velkerri Formation continues to generate substantial interest as an unconventional gas play (Close et al., 2017;Revie, 2017;Baruch et al., 2018); however, the Kyalla Formation also has potential as an oil and gas target (Baruch et al., 2018;Revie and Normington, 2020). Minor oil and gas shows have been identified in wells intersecting Roper Group strata (Table 1) with significant oil shows from the Jamison 1 petroleum well (Jamison sandstone reservoir), that are likely sourced from the Kyalla Formation (Lannigan and Torkington, 1991;Jarrett et al., 2019c;). Additionally, oil bleeds in Urapunga 4 are self-sourced from the Velkerri Formation (Jackson et al., 1986). Significant gas shows in Shenandoah 1A and recent hydraulic fracturing of the Velkerri Formation in Amungee NW-1H confirmed substantial gas flows are possible (Close et al., 2017). Recent resource assessments of the Kyalla and Velkerri formations estimates an in-place resource of 429 million barrels of oil (MMbbls) and between 8 and 118 trillion cubic feet of gas (Tcf) (Revie, 2017;Schneck et al., 2019). Such estimates attest to the untapped economic potential of the Mesoproterozoic Kyalla and Velkerri formations.
In this contribution, we synthesise previous work with newly acquired data focussing principally on the biogeochemical (e.g. organic chemistry, carbon-sulfur systematics, basin redox, organic carbon isotopes) and source rock characteristics (e.g. TOC, thermal maturity) of the Velkerri Formation. Although Paleoproterozoic hydrocarbon plays in the McArthur Basin are currently in the early stages of exploration (e.g. the ca. 1.74 Ga Wollogorang Formation, and the ca. 1.64 Ga Barney Creek Formation, with its correlative, the Riversleigh Siltstone in the Mount Isa Province of Queensland), globally, the Velkerri Formation is one of the most advanced Precambrian unconventional hydrocarbon plays. Considering its Mesoproterozoic age (ca. 1.40-1.34 Ga), the Velkerri Formation offers crucial insights into Precambrian paleoenvironments that may be conducive to exploring for hydrocarbon resources of this antiquity.

Geology
The sedimentary successions of the greater McArthur Basin were subdivided by Rawlings (1999) and Close (2014) into five basin-wide, disconformity-or unconformitybounded depositional 'packages'; in ascending stratigraphic order, the Paleoproterozoic Redbank, Goyder and Glyde packages, and the Mesoproterozoic Favenc and Wilton packages. Each package is characterised by similarities in age, stratigraphic position, lithofacies composition, style and composition of volcanism, and basin-fill geometry. The Roper Group ( Fig. 2) represents the Wilton package in the southern McArthur Basin and covers over ~145 000 km 2 of northern Australia in either surface outcrop or the sub-surface (Jackson et al., 1986;Abbott and Sweet, 2000;Munson, 2016). Significant lateral thickness changes occur within the Roper Group, it is thinnest (<500 m) over the north-south trending Batten Fault Zone, ~1-2 km thick in the vicinity of the Urapunga Fault Zone and thickens dramatically to >5km south of the Urapunga Fault Zone in the eastern Beetaloo Sub-basin (Plumb and Roberts, 1992;Abbott and Sweet, 2000;Rawlings et al., 2004;Munson, 2016;Williams, 2019) (Fig. 1). This has been interpreted to represent the main depocentre of the Roper Group (Plumb and Wellman, 1987;Abbott and Sweet, 2000;Williams, 2019).
Recent palaeomagnetic-based reconstructions Cawood et al., 2020) place the North Australian Craton in mid to low latitudes at ca. 1.3 Ga, plausibly juxtaposed against Laurentia (North America), Siberia, South China and the North China Craton (Fig. 3). Collins et al. (2019) suggested that the broader marine environment formed a large gulf, hereafter referred to as the McArthur-Yanliao Gulf. Both the McArthur Basin and the Yanliao Basin (North China Craton) preserve coeval black shales. The black shales of the Velkerri Formation (northern Australia) are constrained by a combination of absolute ages (Re−Os) and maximum depositional ages (U−Pb detrital zircons) to be between 1.39 ± 0.01 Ga (2σ) and 1.36 ± 0.02 Ga (2σ) (Bodorkos et al., 2020;Kendall et al., 2009;Yang et al., 2018).
The black shales of the Xiamaling Formation, within the North China Craton, were deposited at 1.38 ± 0.01 Ga (Zhang et al., 2015;Zhang et al., 2017), and have comparable TOC levels to that of the Velkerri Formation.
The Velkerri Formation is the initial relatively deep-water facies of a shoaling upward sequence (the interpreted Veloak Sequence of Abbott and Sweet, 2000), which comprises the dominantly basinal facies of the Velkerri Formation transitioning up-section to cross-bedded sandstones of the Moroak Sandstone (Fig. 2). The Velkerri Formation is formally divided into three distinct lithostratigraphic members (Munson and Revie, 2018), the Kalala Member, Amungee Member and Wyworrie Member (oldest to youngest, Fig. 2). The Kalala Member has been interpreted as a highly condensed transgressive system tract. The rest of the Velkerri Formation consists of a high stand systems tract, whose top is the base of the overlying Moroak Sandstone (Abbott and Sweet, 2000). The Amungee Member comprises the primary black shale facies, with total organic carbon contents reaching ~10 wt% (Cox et al., 2016;Revie and Normington, 2020)(Figs. 4 & 5) and has been interpreted by astrochronological methods to have been deposited over ca. 10 million years (Mitchell et al., 2020).
Detailed sedimentary petrology of the Amungee Member has shown that the black shale facies was deposited below storm-weather wave-base and is composed of thinly laminated, gray-green to dark gray clays, pale gray silts and rare, fine-grained sands (Warren et al., 1998;Munson, 2016;Munson and Revie, 2018;Sheridan et al., 2018). Organic enrichment within the Amungee Member occurs principally as three prominent horizons informally referred to as the A, B and C organofacies (oldest to youngest) (Figs. 2,4 & 5). The organic-rich and organic-lean intervals of the Amungee Member show no systematic change in grain size that correlate with TOC enrichment (Warren et al., 1998;Cox et al., 2016). Based on this observation, Warren et al. (1998) noted that these intervals do not reflect changes in the energy of deposition or water depth, but rather changing water column biochemistry.
Systematic changes in nitrogen isotopes corresponding to the organofacies suggests that basin redox and bioavailable nitrogen (the proximal limiting nutrient in most marine systems; Tyrrell, 1999;Bristow et al., 2017), are primary controls on the development of the organofacies , whereas the enrichment of the Amungee Member as a whole has been related to enhanced phosphorus supply (Cox et al., 2016).

Basin provenance
The Velkerri Formation is sandwiched by two compositionally mature quartz arenites, the underlying Bessie Creek Sandstone and the overlying Moroak Sandstone (Fig. 2). The U-Pb and Hf isotopic composition of detrital zircons within the upper Roper Group show a provenance change from predominately 'eastern sources' to 'southern sources' (present-day direction) up stratigraphy Yang et al., 2019) (Fig. 6). Variation in zircon sourcing broadly reflects an excursion to juvenile neodymium isotope compositions in the Amungee Member shales (Cox et al., 2016). Yang et al. (2018) suggested that this is a reflection of the development of an active margin on the southern side of the North Australian Craton at ca. 1.4 Ga as the Mirning Ocean was subducted, providing nutrients to the McArthur-Yanliao Gulf.

Ecology
The indigenous hydrocarbon biomarker assemblage for the Velkerri Formation has been characterised by Jarrett et al. (2019a), who described a water column dominated by bacteria with large-scale heterotrophic reworking of the organic matter in the water column or bottom sediment. Evidence for microbial reworking includes a large unresolved complex mixture (UCM) and high ratios of monomethyl alkanes relative to n-alkanes-features characteristic of indigenous Proterozoic bitumen (Pawlowska et al., 2013). Steranes, biomarkers for single-celled and multicellular eukaryotes (e.g. Peters et al., 2005;Summons et al., 2006), were below detection limits in all extracts analysed, despite eukaryotic microfossils having been previously identified in the Roper Group (Javaux et al., 2004). This work suggests that eukaryotes, while present, were ecologically restricted and contributed little to the net biomass. This was broadly consistent with preliminary results presented by Flannery and George (2014). The combination of increased dibenzothiophene in the Amungee Member and low concentrations of 2,3,6-trimethyl aryl isoprenoids throughout the Velkerri Formation suggests that the water column at the time of deposition was only transiently euxinic and did not extend into the photic zone. Jarrett et al. (2019a) noted that the biomarker assemblages and water column chemistry of the Velkerri Formation differ markedly from the underlying ca. 1.64 Ga Barney Creek Formation (Brocks et al., 2005;Nettersheim, 2017), demonstrating that microbial environments and water column redox were variable in the Proterozoic and may reflect different depositional environments.

Basin redox
Widely used proxies for basin redox include the redox-sensitive trace metals of cerium, molybdenum and vanadium in black shales (Fig. 7). These metals are useful indicators as they are found in low concentrations (i.e. depleted) under oxidising conditions, but are enriched in sediments under reducing conditions. For example, cerium shows relative depletion compared to its neighbouring rare earth elements (i.e. Ce anomaly) due to its incorporation into ferromanganese nodules/crusts (Elderfield and Greaves, 1981;Nath et al., 1994) upon oxidation of Ce 3+ to Ce 4+ . In an analogous fashion, molybdenum exists as the oxidised molybdate ion in seawater (MoO4 2-) but reacts strongly with hydrogen sulfide, such that it is effectively removed from seawater and pore waters under sulfidic (i.e. euxinic) conditions by incorporation into the sediment. Furthermore, molybdenum also complexes with organic molecules; however, this process still requires hydrogen sulfide of ~10 uM (Erickson and Helz, 2000) to quantitatively form particle reactive thiomolybdates (Scott and Lyons, 2012). Similar to molybdenum, vanadium occurs as oxidised V 5+ as vanadate oxyanions (VO4 3-) under oxic conditions, whereas V 5+ is reduced by both organic compounds and hydrogen sulfide to V 4+ under anoxic conditions (Breit and Wanty, 1991), with strong affinities towards organo-metallic complexes (Algeo and Maynard, 2004) and authigenic clays (Peacor et al., 2000).
Due the differing redox potentials of molybdenum and vanadium and coupled redox reactions (i.e. sulfate reduction), V/Mo ratios in shales exhibit distinct changes through the transition from oxic, suboxic, anoxic to euxinic conditions (Piper and Calvert, 2009) (Fig. 7c).
Broadly, V/Mo ratios increase through the oxic to suboxic transition while displaying nearly invariant molybdenum concentrations. However, the suboxic to anoxic and then euxinic transition is marked by decreasing V/Mo ratios with increasing molybdenum concentrations.
To date, three prominent basin redox studies have been conducted on the Velkerri Formation, along with data from other Roper Group formations, to suggest a redox structure for the basin. Shen et al. (2003) described deep water anoxia along with oxygenated shallow waters based on iron speciation data, whereas both Cox et al. (2016) and Mukherjee and Large (2016), used trace metals in shales and pyrite respectively, to argue for oxygenated shallow waters, suboxic to anoxic deep waters and intermittent euxinia.
Here we extend these studies by incorporating new data from five wells that form a broad E-W transect across the basin (Tarlee S3, Altree 2, Amungee NW-1H, Tanumbirini 1, Marmbulligan-1; Fig. 1). These data show that redox stratification is persistent across the basin (Fig. 7). In agreement with previous studies, this redox stratification consists of a shallow oxic layer (Fig. 7 A-B) overlying suboxic to anoxic waters ( Fig. 7 C-D). Euxinia is recorded in all five wells and is exclusively associated with high degrees of organic carbon export during deposition of the Amungee Member (Figs. 4, 5, 7 C-D).
These dynamic redox stratified waters are consistent with the prevailing view that Mesoproterozoic basins were characterised by anoxic (ferruginous) to suboxic deep waters, episodic euxinic continental shelfs and weakly oxic surface waters (Planavsky et al., 2011;Poulton and Canfield, 2011).
From a source rock perspective, this overall period of widespread suboxic to anoxic relatively deep water implies that deep basins were always conducive to the preservation of organic matter, consequently basin redox fails to explain the onset of organic matter enrichment during deposition of the Amungee Member.

Basin salinity and restriction
Understanding basin dynamics, in particular salinity and restriction is difficult in Precambrian successions, where the full stratigraphic record of the basin is not preserved, as in the case of the Velkerri Formation and broader Roper Group. Donnelly and Crick (1988) inferred deposition of Roper Group sediments within a large lake or silled basin based upon isotopically heavy sulfur values (δ 34 S +3.6‰ to +34.4‰) from disseminated pyrite, inferring a low sulfate environment. A marine origin was proposed based on the wide lateral continuity of Roper Basin sediments (Rawlings, 1999;Rawlings et al., 2004), facies associations and stacking patterns typical for sequence development on a siliciclastic continental shelf, and preserved sedimentary structures typical of open marine deposition (e.g. hummocky and swaley cross-stratification) (Abbott and Sweet, 2000). However, Munson (2016) noted the common occurrence of synaeresis cracks in the Velkerri Formation suggesting that water salinity fluctuated during deposition (Plummer and Gostin, 1981). Considering its intracontinental setting, this sedimentological evidence for changes in salinity and/or connectivity suggests a dynamic interplay between marine/fresh water influx, likely facilitated by tectonically-induced basin restriction/connectivity. Sulfur in sediments is largely governed by the availability of seawater sulfate, which is coupled to salinity (Berner and Raiswell, 1983;1984). Carbon-sulfur ratios are sensitive to distinguish between high-and low-sulfate environments and, to a lesser extent, oxic through euxinic settings. These have been validated in both modern (e.g. Berner, 1984;Berner and Raiswell, 1984) and ancient (e.g. Lyons et al., 2000) sedimentary systems. Generally, the supply of sulfate in normal marine settings is not limited, and as long as sufficient organic matter is available for bacterial sulfate reduction, sediments in marine settings are characterised by low C/S ratios. In contrast, freshwater conditions, which favour low dissolved sulfate levels, produce sediments with high C/S ratios (Berner and Raiswell, 1984).
Complications involving C/S ratios exist and include diagenetic, catagenic and metamorphic alteration of both carbon and sulfur, along with the understanding that the reactivity of bulk carbon was likely higher in pre-Silurian sediments due to the absence of lower reactive terrestrial carbon sources (i.e. plant matter, Raiswell and Berner, 1986;Raiswell and Albiatty, 1989;Hieshima and Pratt, 1991), resulting in lower C/S ratios of pre-Silurian sediments.
Carbon and sulfur values from the Velkerri Formation show considerable scatter (Fig.   8), with the majority of data having C/S ratios less than 2.8 (average 2.4 ± 0.5; 2σ CI), similar to modern normal marine waters and an order of magnitude lower than the modern freshwater array (Berner, 1984;Berner and Raiswell, 1984).
Due to the greater reactivity of Precambrian organic matter, Lyons et al. (2000) argued that normal open marine Precambrian C/S ratios should be lower than post-Silurian averages (C/S ~ 1.4) and at least equal to (or lower than) Cambrian ratios (C/S ~ 0.5). Excluding a single sample, all Velkerri Formation values have C/S ratios greater than Cambrian ratios, with most values higher than post-Silurian averages. These results preclude a freshwater origin, but, considering the scatter in the dataset, the possibility of enriched sulfur contents due to euxinic samples, and the high salinities recorded in formational (possibly connate) waters from the above lying Moroak Sandstone (Lannigan and Torkington, 1991), salinities somewhere between brackish to normal marine are likely.

Organic carbon isotopes
High total organic carbon (TOC) contents within sediments have been attributed to various factors. These include high primary productivity (Pedersen and Calvert, 1990;Calvert et al., 1996), high nutrient fluxes due to warm and wet climatic conditions (Condie et al., 2001;Meyer and Kump, 2008), basin redox conditions facilitating the preservation of organic matter (Calvert et al., 1996;Hartnett et al., 1998), mineralogical controls on organic carbon export (Mayer, 1994;Hedges and Keil, 1995;Kennedy et al., 2014), and changes in the relative rate of clastic to biogenic sedimentation (Muller and Suess, 1979;Suess, 1980). These processes, which may vary in both space and time, are important in constraining the development of hydrocarbon source rocks.

Measured organic carbon isotope ratios ( 13 Corg) from samples of the organic-rich
Amungee Member vary between ~-35‰ and ~-32‰. They exhibit a positive correlation with TOC ( Fig. 9), such that the most isotopically heavy 13 C values are associated with the highest degrees of TOC enrichment. Such a relationship suggests that variations in the siliciclastic flux are not controlling carbon enrichment, as this mechanism would not have an isotopic effect.
Also, such organic carbon enrichment is not associated with a maximum flooding surface (Warren et al., 1998;Cox et al., 2019). Considering this, it is likely that the observed 13 C variations reflect changes in dissolved inorganic carbon (DIC), or, more specifically, variations in contemporaneous organic carbon burial (ƒorg), which likely reflects the strengthening of primary productivity. Assuming that the observed variations in 13 Corg reflect changes in a local DIC reservoir, these could be then translated to changes in the fraction of organic carbon buried (ƒorg) in the McArthur-Yanliao Gulf. The latter can be quantified using a simple carbon isotope mass balance approach for the calculation of ƒorg in the oceans (see Kump et al., 2010).
Using modern parameterisation of the marine carbon cycle (i.e. carbon isotope composition of volcanic input of -5‰ to -8‰, Javoy et al., 1986), the fractionation factor between inorganic and organic carbon of ~ 23‰ to 34‰ (Hayes et al., 1999), and a Monte Carlo simulation approach (see Supplementary Information), data suggest a 3-fold increase in ƒorg to produce the observed change in 13 Corg.

Source rock characteristics
Using a large data compilation, including the collation in Revie and Normington (2020) and newly generated data in this study (n=1861), bootstrapped averages and confidence intervals were calculated on all available pyrolysis data to present an overview of the source rock characteristics of black shales associated with the Velkerri Formation. To assess sample bias, the data were "jackknifed" (Tukey, 1958) to ensure no individual value(s) significantly biased the data. In all cases, bootstrapped and jackknifed averages are within 7% agreement (or better) of each other (see Supplementary Information). The whole-rock pyrolysis data (n = 1442) were filtered to present a more constrained view of source rock characteristics based on the criteria discussed in Hall et al. (2016) and in the supplementary information (Fig. 10).
Pyrolysis results may be unreliable for organic-lean samples, or due to the presence of nonindigenous free hydrocarbons-either migrated hydrocarbons or drilling fluid contaminants (Peters, 1986;Dembicki, 2009;Carvajal-Ortiz and Gentzis, 2015). In Amungee NW-1H for example, newly generated Tmax values average 320°C (Table 2). These immature values are inconsistent with mud-gas indications in the same intervals that range from 0.29% to 0.67% (Origin Energy Resources Ltd, 2015). Low S2 values (<0.2 mg HC/g rock) and HI (< 7 mg HC/g rock) are consistent with overmature source rocks and these values were removed during filtering.
Analysis of the black shale facies reveals maximum TOC of ~10 wt%, with an average of 2.65 wt% (± 0.11; 2σ CI); TOC varies between wells (Table 2) (Figs. 4 & 5). The classification scheme published by Peters & Casa (1994) categorized rocks with TOC between 1-2 wt% as good source rocks, 2-4 wt% as very good, and > 4 wt% as excellent. Using kernel density estimates to construct the density distribution, Velkerri Formation source rocks should be considered as good to excellent, with most samples considered as good to very good (Fig. 11).
The source rock units where TOC >2 wt% typically exceed tens of metres in thickness and can be in excess of 200 metres (e.g. Altree 2, Cox et al., 2016) (Fig. 5). The presence of thick intervals of organic-rich shale is a key aspect of the prospectivity of the Velkerri Formation in the Beetaloo Sub-basin. Productive shale-gas systems generally have TOC >2 wt% and are in excess of 45 m thick (Jarvie, 2012).
Velkerri Formation kerogens contain potential to generate both oil and gas, based on present-day hydrogen indices (HI) of up to 800 mg HC/g TOC (Fig. 11). The average presentday HI is 200 mg HC/g TOC (± 8.4; 2σ CI), suggestive of either a gas-prone source rock, or thermally mature sediments (Peters and Cassa, 1994). Results from the thermally immature Marmbulligan well analysed in this study have HI values reaching 551 mg HC/g TOC consistent with an oil-and gas-prone kerogen (Supplementary Information). Trends in hydrogen and oxygen indices are indicative of Type I/II kerogen (Fig. 10 A & D), broadly consistent with previously published results suggesting a bacteria-dominated biomass (Jarrett et al., 2019c;Revie, 2017). Oxygen indices indicate some degradation of organic matter via oxidative processes.
Thermal maturity, as viewed through Tmax values ranges from immature to overmature, with an average of 438°C (± 1.6; 2σ CI) corresponding to relatively low thermal maturity (i.e. oil window) (Fig. 10). Revie (2017) spatially mapped thermal maturity of the Velkerri Formation through the Beetaloo Sub-basin, demonstrating that thermally mature to overmature sediments were situated within the deepest and thickest intersections in the central part of the region where well penetration is limited. Individual well averages diverge from the overall sample population averages. Such divergence is unrelated to sampling density (see Supplementary Information) and likely reflects spatial differences in thermal maturity and original TOC; the former is supported by a statistically significant relationship between HI, S1, Production Index (PI) and depth to intercept (Fig. 12).
Average Tmax values suggest maturation principally in the oil window, whereas the average present-day HI places maturation in the gas window (Fig. 10). However, Tmax can be influenced by TOC content, kerogen type and mineral matrix effects (e.g. Espitalie et al., 1980;Peters, 1986;Dembicki, 2009). An alternative proxy for maturity is the reflectance of organic matter (Dow and O'Connor, 1982). Thermal maturity trends based on both Tmax and organic reflectance show broad correlations down core in four of the wells analysed in this study (Fig.   13). The discordance between apparently reliable maturity proxies in Velkerri source rocks have been discussed in other studies Faiz et al., 2016;Jarrett et al., 2019c) and suggest that the typical Tmax thresholds associated with the oil and gas windows may need refining for Precambrian organic matter. We speculate that the overall higher reactivity of bulk Precambrian biomass (Raiswell and Berner, 1986;Raiswell and Albiatty, 1989;Hieshima and Pratt, 1991), and specifically biomass with no terrestrial matter (Tegelaar and Noble, 1994), may be a significant factor for this discordance; this requires further research. To summarize, thick intervals of organic-rich shale occurring in both the oil-and gaswindows are key aspects of the prospectivity of the Velkerri Formation in the Beetaloo Subbasin. Productive shale-gas systems globally generally have TOC >2 wt%, are in excess of 45 m thick and are within the gas window (Jarvie, 2012). Therefore, the Velkerri Formation meets many of the criteria required for a successful shale-gas play.

Broader earth system implications
Whereas the deposition of organic-rich shales has clear implications for hydrocarbon prospectivity, the geochemistry of the Velkerri Formation has global implications for basin and potentially atmospheric oxygenation, as net oxygen production is fundamentally tied to the burial of organic matter.

The widespread and coeval deposition of black shales in both the McArthur and North
Chinese Yanliao basins at ca. 1.38 Ga (Kendall et al., 2006;Zhang et al., 2015;Zhang et al., 2017;Mitchell et al., 2020), has been used by a number of studies that propose an oxygenation event at this time (Cox et al., 2016;Gilleaudeau et al., 2016;Mukherjee and Large, 2016;Zhang et al., 2016;Zhang et al., 2018;Mukherjee et al., 2019).
Whether such an oxygenation event was sustained or transient, basin-scale or atmosphericscale, is a matter of significant debate Planavsky et al., 2018). The concentration of molybdenum in middle Velkerri shales is one of the most convincing lines of evidence for this (possibly transient) oxygenation. Oxidised molybdenum is soluble in oxic waters; however, under eunixic conditions, it is quickly lost from the water column (Reinhard et al., 2013). Consequently, molybdenum concentrations in shales are tied to the oceanic inventory of dissolved molybdenum and more broadly, to both basin and atmospheric redox (Algeo and Lyons, 2006;Scott and Lyons, 2012). Both initial and more recent compilations of molybdenum concentrations reveal that molybdenum rarely exceeds 50 ppm in the Paleoproterozoic or Mesoproterozoic (Scott et al., 2008;. Velkerri Formation molybdenum concentrations average 59 ppm (±4; 2σ CI), with a maximum of 118 ppm, which suggests that euxinia was not ubiquitous during this period. Consequently, the dissolved molybdenum inventory may have been unusually high at this time in the McArthur-Yanliao Basin, suggesting, at the very least, transiently higher basin and possibly atmospheric oxygen levels.

Implications for hydrocarbon exploration
In many ways, the greater McArthur Basin is not exceptional. Proterozoic basins cover many cratonic areas. For example, the extensive 'Purana' basins of India are similar in terms of age and rock types to the greater McArthur Basin (e.g. Collins et al., 2015). Most of these basins are extremely under-explored.
The example of the Velkerri Formation suggests that high organic horizons correlate with proxies for increased oxia in upper parts of the water column. This introduces an interesting feedback. Increased nutrient supply promotes biomass, which photosynthesises and increases oxygen levels in the upper parts of the ocean. Increased oxia in surface waters means that more oxidised and soluble nutrients such as nitrate (NO3 -) and phosphate (PO4 3-) remain dissolved in these waters, as do many key nutrients and bio-essential trace metals or micro-nutrients (e.g. Fe, Zn, Cu, Mo, Ni) that are highly redox sensitive. In fact, in an atmosphere with about 1% of the partial pressure of O2 of today-similar to estimates of the ca. 1500 Ma Earth-marine oxia would occur only in local 'oxygen oases' where life blooms (Reinhard et al., 2013). Therefore, source rock quality would be expected to vary considerably along strike within a basin due to variations in nutrient supply.

Conclusions
The Mesoproterozoic Velkerri Formation offers unique insights into Proterozoic paleoenvironments and their hydrocarbon potential. Redox stratification was persistent across the basin and consisted of a shallow oxic layer with deeper water suboxic to anoxic conditions that were episodically euxinic. The biomass was dominated by bacteria and multiple lines of evidence suggest that organic enrichment is related to high degrees of primary productivity associated with conducive biogeochemical conditions. Although deposited in an intra-continental setting, the sedimentology, paleosalinity and palaeoredox conditions of the basin support global reconstructions that have the Velkerri Formation lying within a large embayment of the supercontinent Nuna, that we call the McArthur-Yanliao Gulf. These reconstructions, combined with available age constraints, tie the deposition of the Velkerri Formation with a period of more widespread deposition of black shales within this gulf that may reflect high biological productivity that endogenously oxygenated surface waters in a low atmospheric oxygen world. This may reflect local basin conditions in a Baltic Sea-like environment (e.g. Markus Meier et al., 2021) Cox et al. (2016). Dark green circles = Altree 2, red circles = Amungee NW-1H, blue circles = Tarlee S3, light green circles = Tanumbirini 1, yellow circles = Marmbulligan 1. Figure 8. Sulfur-carbon systematics for the Velkerri Formation. Freshwater and Holocene normal marine C/S ratios are from Berner and Raiswell (1983). Post Silurian and Cambrian C/S ratios are from Lyons et al. (2000). Dark green circles = Altree 2, red circles = Amungee NW1, blue circles = Tarlee S3, light green circles = Tanumbirini 1, yellow circles = Marmbulligan 1.   Figure 12. Cross-plot of average pyrolysis data versus depth to interception of the Velkerri Formation. Statistically significant correlations exist at the 95% confidence level for (A) S1 (B) Hydrogen Index (C) Production Index, but no statistically significant relationship exists for TOC. Statistical significance is based on a standard 2-tail test with alpha = 0.05 (pS1 = 0.02, pHI = 0.0035, pPI = 0.0035, pTOC = 0.51). Figure 13. Organic maturity vs depth plots for four wells used in this study. Gray shaded areas are the depths of the Velkerri Formation in each hole. Peach color vertical column is a broad maturity window. Blue diamonds = solid bitumen reflectance values (from Revie & Normington, 2020). Pink triangles = calculated vitrinite-equivalent reflectance equivalents calculated from Tmax using the Jarvie (2018)