Microbial assemblage and palaeoenvironmental reconstruction of the 1.38 Ga Velkerri Formation, McArthur Basin, northern Australia

Abstract The ca. 1.38 billion years (Ga) old Roper Group of the McArthur Basin, northern Australia, is one of the most extensive Proterozoic hydrocarbon‐bearing units. Organic‐rich black siltstones from the Velkerri Formation were deposited in a deep‐water sequence and were analysed to determine their organic geochemical (biomarker) signatures, which were used to interpret the microbial diversity and palaeoenvironment of the Roper Seaway. The indigenous hydrocarbon biomarker assemblages describe a water column dominated by bacteria with large‐scale heterotrophic reworking of the organic matter in the water column or bottom sediment. Possible evidence for microbial reworking includes a large unresolved complex mixture (UCM), high ratios of mid‐chained and terminally branched monomethyl alkanes relative to n‐alkanes—features characteristic of indigenous Proterozoic bitumen. Steranes, biomarkers for single‐celled and multicellular eukaryotes, were below detection limits in all extracts analysed, despite eukaryotic microfossils having been previously identified in the Roper Group, albeit largely in organically lean shallower water facies. These data suggest that eukaryotes, while present in the Roper Seaway, were ecologically restricted and contributed little to export production. The 2,3,4‐ and 2,3,6‐trimethyl aryl isoprenoids (TMAI) were absent or in very low concentration in the Velkerri Formation. The low abundance is primary and not caused by thermal destruction. The combination of increased dibenzothiophene in the Amungee Member of the Velkerri Formation and trace metal redox geochemistry suggests that degradation of carotenoids occurred during intermittent oxygen exposure at the sediment–water interface and/or the water column was rarely euxinic in the photic zone and likely only transiently euxinic at depth. A comparison of this work with recently published biomarker and trace elemental studies from other mid‐Proterozoic basins demonstrates that microbial environments, water column geochemistry and basin redox were heterogeneous.


| INTRODUC TI ON
The relationship between ocean chemistry, microbial metabolisms, and the evolution and proliferation of eukaryotes are themes of long-standing interest in Proterozoic geobiology (Anbar & Knoll, 2002;Johnston et al., 2012;Nursall, 1959). The Mesoproterozoic (1.6-1.0 Ga) Earth is characterised by persistently low atmospheric pO 2 Scott et al., 2008;Zhang et al., 2016), pervasively oxygen-depleted oceans (Canfield, 2004;Frei, Gaucher, Poulton, & Canfield, 2009;Lyons & Reinhard, 2009;Poulton, Fralick, & Canfield, 2004) and an apparently stable carbon cycle, as recorded by the carbon isotope ratios of marine carbonates (δ 13 C carb : Kaufman, 1997). Consequently, the Mesoproterozoic has generally been considered a period of palaeoenvironmental stability within the Earth system, comprising the core of the so-called "boring billion" years (1.8-0.8 Ga; Brasier & Lindsay, 1998;Buick, Des Marais, & Knoll, 1995). This, in turn, correlates with the purple ocean hypothesis of Brocks et al., 2005 where C 40 aromatic carotenoid derivatives such as isorenieratane and okenane detected in mid-Proterozoic rocks, in addition to their diagenetic breakdown products 2,3,4and 2,3,6-trimethyl aryl isoprenoids, suggest notable primary productivity by green sulphur bacteria (Chlorobiaceae) and purple sulphur bacteria (Chromatiaceae) where anoxic conditions reached into the photic zone of the water column (e.g., Brocks et al., 2005;Koopmans, Schouten, Kohnen, & Sinninghe Damsté, 1996;Schwark & Frimmel, 2004). Sub-oxic to anoxic shallow waters and euxinic  Ahmad et al. (2013). Note that the Derim Derim dolerite sills (not shown) intrude at multiple levels through the Roper Group from the Mainoru Formation through to the base of the Kyalla Formation [Colour figure can be viewed at wileyonlinelibrary.com] deep-water conditions are considered hostile for the expansion of eukaryotes which are largely aerobic and poisoned by sulphide (Anbar & Knoll, 2002). Such conditions were presumably widespread in the mid-Proterozoic based on biomarker evidence for photic zone euxinia in the 1.64 Ga Barney Creek Formation (Brocks et al., 2005), and the 1.1 Ga Atar Group, Mauritania (Blumenberg, Thiel, Riegel, Kah, & Reitner, 2012;Gueneli, Brocks, & Legendre, 2012), in addition to iron speciation and trace element geochemistry (Canfield, 2004;Poulton et al., 2004) suggested euxinia was a defining feature of the Mesoproterozoic and led to the restriction of eukaryotes to local oases with stable oxygenated conditions. However, in spite of apparent homogeneous and widespread anoxia in deep-water Mesoproterozoic oceans, redox heterogeneities and fluctuations have been identified from the ca. 1.4 Ga Kaltasy Formation, Southern Ural Mountains, Russia, a carbonaceous shale deposited in a basinal, oxic environment (Sperling et al., 2014;Sergeev, Knoll, Vorob'eva, & Sergeeva, 2016); the ca. 1.4 Ga Xiamaling Formation of the North China craton, a siliciclastic shale deposited in a low-energy deep, sub-tidal environment (Canfield, 2014;Diamond, Planavsky, Wang, & Lyons, 2018), and the age equivalent Velkerri Formation (Cox et al., 2016), Roper Group, northern Australia (Figure 1) suggesting that Mesoproterozoic sequences previously studied may have reflected geochemistry of restrictedbasin settings and contained a taphonomic bias that a larger sample set of biomarker, fossil and geochemistry of different depositional environments including nearshore and oxygenated sedimentary basins may rectify.

The geochemical structure of the Roper Seaway in the McArthur
Basin involved a shallow oxic layer overlying deeper waters that were sub-oxic to anoxic with intermittent euxinia in the Amungee Member of the Velkerri Formation based on redox-sensitive trace element data (Figure 2; Cox et al., 2016). Euxinic conditions, as determined by trace element abundances (e.g., enrichment of molybdenum, vanadium and uranium above values typical of the Post Archean Australian Shale), developed once primary productivity was significantly greater than the flux of oxygen, nitrates and oxidised iron, resulting in sulphate reduction (Johnston et al., 2010).
The Roper Group of the McArthur Basin ( Figure 1) is also one of the sites that contain the oldest evidence for fossils that have a clear eukaryotic origin (Grey, 2015;Javaux et al., 2001;Peat, Muir, Plumb, McKirdy, & Norvick, 1978). The shoreline facies of the Corcoran and Jalboi formations show the greatest abundance and diversity of biota including the acritarch Tappania plana which shows cell organisation and evidence for reproduction via budding that are unique to F I G U R E 2 Stratigraphic column of drillcore Altree 2 demonstrating variation in organic and inorganic composition through the Velkerri Formation, A-gamma log (API), B-Total organic carbon (TOC %), C-T max (°C), D-Hydrogen Index (HI = (S2/TOC) × 100, mg HC/g rock), E-Methylphenanthrene distribution factor (MPDF = (3 + 2)/(3 + 2+9 + 1)-methylphenanthrene, integrated in the m/z 192 trace), F-Dibenzothiophene/phenanthrene ratio (DBT/P) integrated in the m/z 184 and 178 traces, G-Iron/aluminium Fe(wt %)/Al(wt %) with a vertical dashed line in the plot at 0.5 to differentiate between oxic and anoxic conditions (Lyons & Severmann, 2006), H-Molybdenum (Mo ppm), shaded area indicates the approximate threshold value for intermittent to persistent euxinia , I-Vanadium (V ppm), J-Pyrite (wt %) ppm. Trace element (Al, Fe, Mo, V) and XRD (pyrite wt%) data are replotted from Cox et al. (2016) (Javaux, Knoll, & Walter, 2001). Increasing diversity towards the shoreline may be due to either the restriction of bioessential trace nutrients to these environs (Anbar & Knoll, 2002) or a taphonomic bias (Javaux et al., 2001). Hydrocarbon biomarkers have a different taphonomic window than fossils (Brocks et al., 2016) and are a complementary proxy to determine whether eukaryotes may have also inhabited distal shelf environments of the Roper Seaway.
Historical studies of biomarkers from the Velkerri Formation focused on its petroleum potential, as the Roper Group contains some of the world's oldest organic-rich source rocks (TOC content up to 10%) which have generated both oil and gas within the Beetaloo Sub-basin of the McArthur Basin (Ahmad, Dunster, & Munson, 2013;Jackson, Powell, Summons, & Sweet, 1986;Jackson, Sweet, & Powell, 1988;Summons, Powell, & Boreham, 1988; Figure 1).  (Brocks, Grosjean, & Logan, 2008;French et al., 2015;Grosjean & Logan, 2007). For example, Brocks (2011) measured the spatial distribution of hydrocarbons in 2.7 Ga old Archean shales from the Pilbara in Western Australia. These ancient shales contained indigenous polyaromatic hydrocarbons. A spatial analysis of hydrocarbons between the drill core interior and the outer rounded surfaces revealed saturated hydrocarbons, including steranes and hopanes, were concentrated towards the outer rounded surface of the drill core. Additionally, bulk analyses of dolomitic shales from the 1.64 Ga Barney Creek Formation consistently yielded hopanes and steranes (Summons et al., 1988). However, interior/exterior experiments have concluded that the bulk of the hopanes were indigenous whilst the steranes were contaminants suggesting that the steranes were introduced during drilling and/or storage .
It has also been shown that fissile shale may be pervasively contaminated within the inner-core and no amount of cleaning or sawing can remove these contaminants (e.g., Brocks, 2011;Jarrett, Schinteie, Hope, & Brocks, 2013). Thus, the distribution of hydrocarbons between the interior and exterior within samples to detect concentration is crucial prior to biomarker interpretation (Brocks, 2011;French et al., 2015;Jarrett et al., 2013;Schinteie & Brocks, 2014 & George, 2014). The study demonstrated that steranes were abundant on the exterior surfaces of the drillcore, though below detection limits in the interior portion, and thus were likely to be contaminants. The small sample size of that study (n = 3) and a relatively high thermal maturity of samples warrants a more detailed investigation into the hydrocarbon composition of the Velkerri Formation, particularly in core going through the oil window. Furthermore, with the recent publication of high-resolution inorganic geochemistry of the Velkerri Formation in the Altree 2 drillcore (Cox et al., 2016), this study provides a complementary contaminant-free analysis of hydrocarbon biomarkers preserved in the ca. 1.38 Ga Roper Seaway sediments in a palaeoenvironmental context.

| Velkerri formation, Roper Group, McArthur Basin
The geology of the Roper Group has been summarised by Abbott and Sweet (2000), Jackson, Muir, and Plumb (1987), and Sweet and Jackson (1986). The Roper Group was deposited in the McArthur Basin, northern Australia, and consists of a ca. 1-to 5-km-thick package of regressive-transgressive cycles of siliciclastic rocks preserved over ca. 145,000 km 2 with the main depocentre located in the Beetaloo Sub-basin (Abbott & Sweet, 2000;Plumb & Wellman, 1987; Figure 1). The deeper water Velkerri and Kyalla formations are inferred to be the major source rocks for the oil and gas in the Beetaloo Sub-basin (Ahmad et al., 2013;Crick, Boreham, Cook, & Powell, 1988;Jackson et al., 1986Jackson et al., , 1988Summons et al., 1988).
Two contrasting tectonic models have been proposed for basin formation and the origin of the Roper Group. One model suggests that the Roper Group was deposited in marine to shelf environments on an epicontinental platform in response to lithostatic extension and sagging (Betts & Giles, 2006;Foster & Ehlers, 1998;Spikings, Foster, Kohn, & Lister, 2001. A second model suggests that the Roper Group was deposited on an intracratonic ramp that developed during orogenic flexure (Abbott & Sweet, 2000). Both models suggest the Velkerri Formation was deposited in a deep-water marine environment.  (Jackson, Sweet, Page, & Bradshaw, 1999) and baddeleyite from the Derim Derim Dolerite (Abbott et al., 2001), and detrital zircon U-Pb (Yang et al., 2018).
Deposition of the Velkerri Formation has been constrained to be- the Oils of Eastern Australia report (Summons et al., 2002). The report contains "OilMod" parameters for all oils and oil stains in Eastern and Central Australia including ratios of n-alkanes to pristane and phytane, concentrations and ratios of sterane and hopane homologues in addition to δ 13 C saturates and δ 13 C aromatics . McArthur Basin oil stains contained low concentrations of steranes and hopanes that were deemed "too low to determine source," and δ 13 C saturates and δ 13 C aromatics values ranging from ~−32‰ to −34‰, which are isotopically depleted compared to most Phanerozoic oils (Summons et al., 2002).

| Samples
In this study, 110 samples from the upper, middle and lower Velkerri Formation of the Altree 2 well were sampled at intervals of approximately 10 m and analysed by Rock-Eval pyrolysis, X-ray fluorescence (XRF) and quadrapole inductively coupled plasma mass spectrometry (Q-ICP-MS; see Cox et al., 2016). Additionally, 16 thermally immature to mature rocks within the late oil window (T max 426-461°C; Figure 2) at approximately 50-m intervals through the Velkerri Formation (Table 1) were selected for organic extraction and analysis by gas chromatography-mass spectrometry (GC-MS).

| ME THODS
The methods for inorganic and organic geochemistry have been described in detail elsewhere (Boreham & Ambrose, 2007;Cox et al., 2016;Jarrett et al., 2013;Lee & Brocks, 2011). TOC measurements and Rock-Eval pyrolysis measurements were determined on powdered rock samples via pyrolysis using a Rock-Eval 6 ™ instrument. Rocks were treated with a rigorous screening technique to determine and remove surficial contamination (e.g., Jarrett et al., 2013). Powdered rock samples were extracted using organic solvents and analysed by GC-MS using a Hewlett Packard HP 6890 gas chromatograph interfaced to an HP 5975 mass spectrometer in full scan mode (Boreham & Ambrose, 2007). Metastable Reaction Monitoring (MRM) analyses were performed using an Agilent 6890 gas chromatograph interfaced to a Micromass Autospec Premier double sector mass spectrometer (Jarrett et al., 2013). All samples analysed by MRM were injected in n-hexane to avoid FeCl 2 build-up in the MS ion source and subsequent deterioration of chromatographic signals through the use of halogenated solvents (Brocks & Hope, 2014).
Ancient rocks can be affected by trace amounts of contamination potentially altering the indigenous hydrocarbon signature Jarrett et al., 2013). Laboratory system blanks covering the sawing, crushing, extraction, fractionation and analysis process contained near-zero hydrocarbon contamination background levels.
Using rigorous assessments for syngeneity as described previously (e.g., , the 16 interior fractions contain indigenous biomarkers. Only, these untainted extracts are investigated further. The δ 13 C of individual n-alkanes were measured by gas chromatography-combustion-isotope ratio mass spectrometry (GC-C-IRMS) using a Thermo Scientific MAT 253 isotope ratio mass spectrometer.
Briefly, the hydrocarbon gas components were chromatographically separated using a DB-5 capillary column (60 m × 0.32 mm ID. Film thickness 0.25 μm; Agilent) with ultra-high-purity (UHP) helium as a carrier gas at a constant flow rate of 2.0 ml/min. The on-column injector temperature was held at 150°C. The 2 μl of pentane solution with sample n-alkanes was injected either manually or with the GC PAL autosampler. The δ 13 C values are reported relative to VPDB. All analyses were run as a minimum in duplicate with a maximum experimental error of 0.5‰.

| Bulk characteristics
The bulk characteristics of the Velkerri Formation in Altree 2 are published by Cox et al. (2016) for all samples plotted in Figure 2.
T max values increase downcore from 427°C at 394.1 m to 485°C at 1,220.65 m ( Figure 2). The onset of oil generation in the McArthur Basin has been calculated to occur at ~435°C and the end of the oil window at 470°C (Crick, 1992;Crick et al., 1988). Therefore, T max values for the upper Velkerri Formation suggest thermal immaturity Altree 2 (Crick, 1992;Warren et al., 1998), has locally increased the thermal maturity. ). An increase in TOC at a dolomitic siltstone layer between 1,117 and 1,127 m correlates with an increase in HI. Variations in HI in Altree 2 appear to be influenced by both initial TOC and thermal maturity ( Figure 2). In the Wyworrie Member and the upper Amungee Member, HI and TOC correlate. After the onset of hydrocarbon generation in the lower Amungee Member, the HI broadly decreases downcore due to expulsion of hydrocarbons (Bordenave, 1993).

| Saturated hydrocarbons
Total ion currents of the saturated hydrocarbon fraction for representative samples from the Wyworrie, Amungee and Kalala members of the Velkerri Formation are illustrated in Figure 3. All samples analysed are characterised by a large unresolved complex mixture (UCM) and a higher than usual proportion of monomethyl alkanes relative to regular n-alkanes (Figure 3), such as is typical of indigenous Proterozoic hydrocarbons (Pawlowska, Butterfield, & Brocks, 2013). All extracts of the Velkerri Formation contain n-alkanes from C 11 to at least C 27 , with a maximum at either C 13 or C 15 , and an oddover-even predominance of carbon numbers in the range C 12 -C 18 .
This odd-carbon numbered predominance is also a typical feature in many Ordovician bitumens and oils, where a plausible source is the phototrophic micro-organism Gloeocapsomorpha prisca (Fowler, 1992 and references therein; Peters, Walters, & Moldowan, 2005).
However, other studies have shown that this feature is generally typical in cyanobacteria (Evans, 1994;Lea-Smith et al., 2015;Peters et al., 2005), which is the most likely source of the n-C 13 , n-C 15 and n-C 17 predominance in the Velkerri Formation.
Monomethyl alkanes were detected in all samples including both mid-chain and 2-methyl and 3-methyl (iso-and anteiso-) terminal branching isomers with similar carbon number distribution as the n-alkanes. This is similar to the work by Summons et al. (1988) who used MRM to characterise the monomethyl alkanes in the Velkerri and Barney Creek formations. Summons et al. (1988) identified that the Velkerri Formation contained mid-chained MMAs dominated in low molecular weight ranges (C 13 -C 19 ), and the terminal branching isomers became predominant with higher molecular weights.
Immature Barney Creek Formation extracts contain high ratios of terminal branching isomers compared to mid-chained isomers, and these reduce to comparable abundances in the more mature samples (Summons et al., 1988). Regularly branched isoprenoids up to
At that depth, the thermal maturity increases to the peak oil window (T max 432°C; Rc(MPDF) 0.70%; Table 2), suggesting that the decreasing hopane abundances downcore in Altree 2 reflect thermal breakdown. The homohopane index (C 35 /∑C 31 -C 35 hopanes) is a measure of hopane sulphurisation. Under oxic conditions, it has been shown that the C 35 hopanepolyol side chain can degrade, generating lower molecular weight hopanes, while sulphurisation will preferentially preserve C 35 homologue (Peters et al., 2005). The homohopane index is <0.15 for all samples analysed which may indicate the absence of sulphurisation in the Wyworrie and Amungee Members (Table 2). However, with the onset of thermal maturity, C 31 is preferentially desulphurised, limiting the proxy to immature sediments (Köster, Van Kaam-Peters, Koopmans, De Leeuw, & Sinninghe Damsté, 1997;Peters et al., 2005). Thus, limiting the utility of the homohopane index in the Amungee Member.
The presence of 3β-methylhopanes may indicate the presence of methanotrophic bacteria (Farrimond, Talbot, Watson, Schulz, & Wilhelms, 2004;Peters et al., 2005). Methanotrophic bacteria often inhabit the boundary between anoxic and oxic conditions and are particularly active in water bodies with low sulphate content (Beal, Claire, & House, 2011;Farrimond et al., 2004;Hanson & Hanson, 1996). Inorganic geochemistry is consistent with this preferred environment for methanotrophs. The water column during deposition of the Wyworrie Member is predominantly sub-oxic to anoxic at depth (Cox et al., 2016).

| Aromatic hydrocarbons
Thermal maturity parameters based on the methylphenanthrene distribution factor (MPDF) and methylphenanthrene ratio (MPR; Radke, Garrigues, & Willsch, 1990;Radke, Willsch, & Leythaeuser, 1982) increase with depth ranging from Rc (MPDF) = 0.54%-1.27% (Table 2). These values correspond to the onset of the oil window (Rc ~0.6%) to the late oil window (Rc ~1.8%) and are similar to thermal maturities based on aromatic hydrocarbons in Altree 2 (Summons, Taylor, & Boreham, 1994) and in other wells through the Velkerri Formation Warren et al., 1998;George & Ahmed, 2002). George & Ahmed (2002) undertook a comprehensive study of 16 different aromatic maturity parameters in three wells in the Velkerri Formation. The methylphenanthrenes were shown to be the most sensitive to maturity variations from thermally immature to late oil window (George & Ahmed, 2002). This provides confidence in the thermal maturity results in this study. Additionally, the aromatic thermal maturity values correlate with T max values ranging from 438°C, onset of oil generation, to 461°C towards the end of the oil window (Table 2; Figure 2).
Additionally, there are two clusters of DBT/P ratios in the extracts; the first cluster, comprising Kalala and Wyworrie Member, contains low relative abundances of S and low DBT/P ratios (Figure 2). The Amungee Member has higher S contents and elevated DBT/P ratios probably due to the incorporation of reduced sulphur (H 2 S) into the organic matter. Increases in DBT have been shown to broadly correlate with increasing sulphide concentrations in the deep-water column that is then in turn incorporated into the organic matter (Hughes et al., 1995;Werne et al., 2008). Therefore, these aromatic hydrocarbons may provide evidence for deep-water euxinia in the  Adam, Wehrung, & Albrecht, 1977;Grice, Schaeffer, Schwark, & Maxwell, 1996;Koopmans et al., 1996;Brocks et al., 2005). Intact C 40 carotenoids were not detected in any of the Altree 2 extracts. However, aromatic carotenoids are commonly cleaved into shorter fragments during diagenesis, yielding 2,3,4-trimethyl aryl isoprenoids (TMAI) as the breakdown products of okenone and other carotenoids of purple sulphur bacteria, and 2,3,6-TMAI as the cleavage products of chlorobactene and isorenieratene found in green sulphur bacteria (Brocks & Summons, 2003;Koopmans et al., 1996;Schwark & Frimmel, 2004). Figure 7 investigates the presence or absence of 2,3,4-and 2,3,6-TMAI in the Velkerri Formation by juxtaposing m/z 134 partial mass chromatograms of an Altree 2 samples (black trace) with bitumen from the older Barney Creek Formation where these biomarkers are abundant (blue trace; Brocks et al., 2005).
In such comparisons, it is important to recognise that successful identification of TMAI relies on the presence of a continuous pseudohomologous series with plausible relative abundances, and that many of the short-to medium-chain length pseudohomologues (e.g., C 11 -C 18 ) may coelute with trimethylbenzenes (TMAB) and other interfering compounds (Figure 8). For the Altree 2 sample in Figure 7, a complete series of 2,3,6-TMAI appears to be present in the range C 16 -C 22 . However, the concentration of these compounds is low and their identification must remain tentative due to the presence of numerous interfering signals. The Altree 2 sample also has a few peaks at the right elution position for the 2,3,4-TMAI series. However, there is virtually no signal for the C 18 and C 21 pseudohomologs, so the entire series must be interpreted as not detected.
The genuine absence or low concentration of TMAI in the Velkerri Formation is important because it distinguishes its environment from other Proterozoic formations such as the Barney Creek Formation (Brocks et al., 2005) and the 1.1 Ga El Mreïti Group (Gueneli et al., 2018). We therefore investigate whether the extremely low concentrations of TMAI reflect primary absence or are due to thermal breakdown or oxygen exposure during degradation.
To determine whether 2,3,4-and 2,3,6-TMAI were removed from Velkerri Formation bitumens by thermal breakdown, maturity proxies from the Altree 2 drillcore were compared with a Barney Creek Formation profile from drillcore GR 7 (Figure 8). The Barney Creek Formation in the GR 7 drillcore is composed of dolomitic mudstones and siliciclastic shales that range in maturity from immature (T max 430°C at 50.63 m) to the peak oil zone at the base of the drillcore (T max 446°C at 854.30 m; Lee & Brocks, 2011;Revie, 2016). Thermal breakdown downcore of β-carotene, C 40 aromatic carotenoids and TA B L E 3 n-Alkane carbon isotopic composition (δ 13 C in ‰ vs PDB) for Velkerri Formation samples from Altree 2  Figure 8), while TMAI decrease and TMABs become relatively more abundant with increasing depth and thermal maturity (Figure 8). The C 15 2,3,6-TMAI/TMAB ratio decreases from 6.7 at 45.35 m to 0.11 at 869.60 m, and C 18 2,3,6-TMAI/TMAB decreases from 15.1 to 0.36 over the same depth range. The C 15 and C 18 2,3,6-TMAI/TMAB ratios fall below 1 by 426.7 m (T max 434°C) suggesting that in GR7 the ratio reaches unity at the onset of the oil window. The detection of TMAIs is sensitive to thermal maturity, and the presence of phototrophic sulphur bacteria and PZE conditions can presumably only be detected reliably at immature to moderate thermal maturities, at least in Proterozoic organic facies.
Consequently, either the Roper Basin was never inhabited by phototrophic sulphur bacteria because anoxic waters never reached into the photic zone, or the biomarkers of these organisms were destroyed by heterotrophic reworking, most likely through (intermittent) exposure to oxygen at the sediment-water interface (e.g., Repeta, 1989).
Complementary inorganic geochemical proxies suggest that the deep water was sub-oxic to anoxic throughout the Velkerri Formation, with intermittent euxinia occurring only in the Amungee Member (Figures 2 and 9; Cox et al., 2016). Euxinia correlates with increased DBT/P ratios demonstrating enhanced organic matter sulphurisation ( Figure 2). Carotenoids are diagenetically altered by sulphurisation which can enhance preservation (Kohnen, Sinninghe Damste, & De Leeuw, 1991;Payzant, Montgomery, & Strausz, 1986;Schaeffer, Reiss, & Albrecht, 1995). The recorded sub-oxic conditions mean that sub-oxic degradation of any TMAIs remains a possibility in the Velkerri Formation. Secondary destruction of biolipids in the sediment is also consistent with the generally low concentration of other biomarkers such as cheilanthanes and hopanes in Velkerri sediments when compared to total hydrocarbon concentrations. In summary, the data most strongly suggest that C 40 carotenoids and associated 2,3,4-and 2,3,6-trimethyl aryl isoprenoids are not detected as green sulphur bacteria (Chlorobiaceae) and purple sulphur bacteria (Chromatiaceae) were either not significant in the Roper Seaway, or removed due to deepwater sub-oxic conditions and diagenetic removal. This highlights the utility of a combined inorganic and organic geochemical palaeoenvironmental reconstruction.

| Microbial ecology of the Velkerri formation
Biomarkers from drillcore Altree 2 describe an environment dominated by bacteria and little input of biomass from archaea and eukaryotes ( Figure 9). Evidence for this includes an abundance of mid-chained and terminally branched monomethyl alkanes that may be derived from the isomerisation of functionalised bacterial n-alkyl and methylalkyl lipids (e.g., Figure 3; Pawlowska et al., 2013) in addition to regular, rearranged and methyl hopanes (Figures 4 and 5) in the upper half of the Velkerri Formation.

Saturated steranes are below detection limits in all Velkerri
Formation extracts. Possible explanations for their absence include thermal destruction, taphonomic bias through selective heterotrophic removal of sterols (Pawlowska et al., 2013), or primary absence of the source organisms. Figure 2 demonstrates that the Wyworrie and part of the Amungee Member samples are immature to early mature. At these maturities, steranes are not in the zone of thermal destruction (Peters et al., 2005). Furthermore, since hopanes have a similar thermal stability threshold as steranes and are detectable to 793.37 m, the absence of detectable steranes in the upper, immature zone of the drillcore is clearly not caused by thermal destruction. Pawlowska et al. (2013) suggested that the paucity of steranes in many mid-Proterozoic shallow-water environments may be caused by extreme heterotrophic reworking of planktonic eukaryotic organic matter within the upper, oxygen saturated layer of phototrophic microbial mats ("mat-seal hypothesis"). However, the black organic-rich siltstones of the Velkerri Formation were clearly deposited in a relatively deep, sub-storm wave base environment where oxygen-producing benthic mats cannot thrive. Also unlikely is the selective removal of eukaryotic biomass through extensive recycling within the water column as suggested by Pawlowska et al. (2013). Larger, fast sinking eukaryotic particles would have a shorter residence time in the water column than slow-sinking bacterial debris and should be preferentially preserved rather than destroyed (Gueneli et al., 2018). Thus, selective and extreme heterotrophic sterol removal is not a likely scenario for the paucity of steranes in the Velkerri Formation, and the most likely explanation is that modern-day equivalent, sterol-producing aerobic eukaryotes were not a significant component of the pelagic ecosystem. This is consistent with an absence of steranes in the indigenous Velkerri Formation extracts analysed by Flannery and George (2014) F I G U R E 9 A speculative conceptual model of the chemistry of the Roper Seaway through the (a) upper and lower, and (b) middle Velkerri Formation. Red shading indicates ferruginous waters, purple euxinic waters, and blue the upper oxygenated water column; green indicates cyanobacteria inhabiting the upper water column. Dibenzothiophene molecule shown in (b) to identify higher concentrations in euxinic sediments. Water column geochemistry is based on Cox et al. (2016)  from other locations, suggesting the results from Altree 2 are likely spatially similar through the Roper Seaway, and are consistent with a general lack of steranes in all other mid-Proterozoic basins where indigenous biomarkers have been recovered (Gueneli et al., 2018).
Based on the current record, indigenous steranes only emerge in the Tonian (1,000-720 Ma). However, Tonian steranes maintain a primitive homologue distribution with a nearly 100% predominance of cholestane (C 27 ), and traces of ergostane (C 28 ) and the unique sterane cryostane (26-methylcholestane; Adam, Schaeffer, Paulus, & Brocks, 2017;Brocks et al., 2016). Stigmastane (C 29 ) remains undetectable in the Tonian . Sterane diversity and abundance dramatically increase in the Cryogenian (720-635 Ma), in the warm period between the Sturtian and Marinoan Snowball Earth glaciations 659-645 Ma ago, heralding the rise of algae as ecologically important primary producers in the oceans . The siltstones of the Velkerri formation were thus laid down several hundred million years before eukaryotes became a notable component of marine ecosystems, and the paucity of steranes, and presence of hopanes, points to phototrophic bacteria as the dominant primary producers in the Roper seaway.
It has been hypothesised that anoxic and nutrient-limited open waters in the Mesoproterozoic were hostile to the expansion of eukaryotes (Anbar & Knoll, 2002). Eukaryotic microfossils have been identified in shoreline facies of the Roper Group (Corcoran and Jalboi formations) and decrease in species abundance and diversity towards the deeper water Roper Group facies (Javaux et al., 2001).

A transect of bulk nitrogen (δ 15 N bulk ) isotopes through the Roper
Group demonstrated a lateral gradient in nitrate availability with an increase towards shallow nearshore environments (Koehler, Stüeken, Kipp, Buick, & Knoll, 2016). This may help explain the microfossil distributions (Javaux et al., 2001) in the Roper Group with the abundance of microfossils recorded in nearshore environments in the Roper Group being due to the restriction of nutrients rather than taphonomic bias. It is also consistent with the absence of steranes in the deep-water facies of the Roper Group (this and previous biomarker studies). However, no biomarkers have been reported to date from more oxygenated, fossiliferous shallow-water facies because these facies are commonly organically too lean to yield hydrocarbons. The search for eukaryotic biomarkers in Mesoproterzoic shallow-water facies is subject of future research.
In source rocks, the UCM is commonly attributed to either reworking of the organic matter by heterotrophic micro-organisms either in the water column, or within a microbial mat (e.g., Craig et al. 2013;Li, Cao, Hu, & Luo, 2015). Similar modern environments with microbially derived UCMs include sediments from Lizard Island on the Great Barrier Reef, Australia and St Croix, US Virgin Islands (Reitner et al. 2000), and iron-rich biofilms in Sweden (Heim, Quéric, Ionescu, Schäfer, & Reitner, 2017). A GCxGC TOFMS study on a UCM extracted from the Mesoproterozoic Xiamaling shale in China proposed two causes of the UCM: microbial activity and enhanced isomerisation. In that study, the presence of 1-3 cyclic paraffins and alkanes with multiple branching positions was interpreted as evidence for intensive isomerisation post-burial (Li, Cao, Hu, & Luo, 2015;Mißbach et al., 2018). This is also a possibility for the Velkerri Formation as the abundance of cyclic hydrocarbons and MMA and DMA relative to n-alkanes is high in comparison to typical Phanerozoic bitumens. The causes for such enhanced isomerisation in Proterozoic sediments, however, remain obscure.
The very low relative abundances of pristane and phytane, and the absence of both crocetane and the C 20+ acyclic isoprenoids, demonstrate that archaea were not dominant contributors to the biomass either (Hoffmann, Foster, Powell, & Summons, 1987;Kenig et al., 1995;Peters et al., 2005). By default, bacterial photo-and heterotrophs must then have been responsible for the bulk of export production during deposition of the Velkerri Formation.
An increase in DBT/P occurs within the Amungee Member of the Velkerri Formation. Increases in DBT/P are caused by the reaction and incorporation of sulphur species into organic matter during early diagenesis (Hughes et al., 1995;Werne et al., 2008). The absence of the 2,3,4-and low concentrations of the 2,3,6-trimethyl aryl isoprenoids compared with the Barney Creek Formation (Figure 8) suggests that the photic zone of the Roper Seaway may not have been persistently euxinic and may be in direct contrast with the increased DBT/P ratio. Cox et al. (2016) demonstrated Altree 2 contains elevated concentrations of the elements sulphur, molybdenum and vanadium, recording the onset of intermittent deep-water euxinia, and this is coeval with the increase in DBT/P (Figures 2 and 9; Cox et al., 2016). However, this deep-water euxinia was intermittent with prevailing conditions characterised by oxic shallow waters and sub-oxic deep waters; hence, oxic degradation of aromatic carotenoids derivatives is a distinct possibility in the Velkerri Formation including the degradation of aryl isoprenoids and carotenoids and also lower homohopane indices (Table 2). Therefore, the combination of organic and inorganic biomarker proxies leads to an increased understanding of the microbial ecology of the Roper Seaway in the Mesoproterozoic. 2,3,4-and 2,3,6-TMAI assigned to phototrophic purple sulphur bacteria (of the family Chromatiaceae) and green sulphur bacteria (Chlorobiaceae) respectively (Brocks & Schaeffer, 2008;Brocks et al., 2005), β-carotene presumably derived from cyanobacteria (Lee & Brocks, 2011) and oligoprenyl-curcumanes that may play a role in the oxidative stress response in bacteria (Brocks, Bosak, & Pearson, 2009). However, as in the Velkerri, diagnostic eukaryotic steranes were not detected (Figure 7 in . These biomarkers indicate phototrophic bacterial activity in the oxygenated mixed zone and in underlying anoxic waters within the photic zone of the water column. Independent inorganic geochemistry including iron speciation and sulphur isotope systematics supports euxinic conditions in the BCF (Johnston et al., 2008;Shen, Canfield, & Knoll, 2002). These geochemical conditions and biomarker assemblages are broadly comparable to black, finely laminated shales from the ca. 1.1 Ga El Mreïti Group in the Taoudeni Basin, Mauritania (Blumenberg et al., 2012;Gueneli et al., 2018). There, the presence of 2,3,4-and 2,3,6-TMAI, in addition to iron speciation, and sulphur isotopes of pyrite and redox-sensitive trace elements all suggest nearshore PZE conditions (Blumenberg et al., 2012;Gilleaudeau & Kah, 2015;Gueneli et al., 2018;Kah, Bartley, & Teal, 2012).

| Comparison of the Roper Seaway with other Mesoproterozoic environments
Moreover, the nitrogen isotopic composition of porphyrins from the El Mreïti Group confirms that primary productivity was dominated by bacteria (Gueneli et al., 2018), and there is no fossil (Beghin, Guilbaud, et al., 2017; or biomarker evidence for the activity of algae (Blumenberg et al., 2012;Gueneli et al., 2018). Thus, the combination of these studies, in addition to trace metal redox work in other locations, suggests environmental heterogeneity in the Mesoproterozoic (e.g., Canfield, 1998;Lyons, Anbar, Severmann, Scott, & Gill, 2009).
In those studies, water column chemistry was reconstructed using redox-sensitive trace metals and biomarkers, including the absence of aryl isoprenoids, reducing the probability of photic zone euxinia (Luo et al., , 2016. Furthermore, the ca. 1.1 Ga lacustrine Nonesuch Formation, USA, is an additional example of a ferruginous water body based on extensive Fe-C-S systematics (Cumming et al., 2013). Aryl isoprenoids are also below detection limits in the Nonesuch Formation (Imbus, Engel, Elmore, & Zumberge, 1988;Pratt et al., 1991), strengthening the argument against PZE, although the absence of these hydrocarbons may also be due to other factors including thermal destruction. In addition, a persistently oxic water column in the 1.4 Ga Kaltasy Formation in the Ural Mountains, central Russia has been identified based on iron speciation (Sperling et al., 2014).
The biomarker assemblage of the Velkerri Formation is similar to those reported for the Hongshuizhang, Xiamaling and Additionally, the Xiamaling contains trace concentrations of diahopanes and increased concentration of dibenzofuran interpreted to be biomarkers for aerobic diagenesis . Luo et al. (2016) suggest that the microbial input between these three formations could be varied; however, it is also likely that variations in hydrocarbon distributions are influenced by processes in the water column and bottom sediments including clay-catalysed isomerisation reactions (e.g., Alexander, Kagi, & Larcher, 1984), the amount of sulphurisation (Schaeffer et al., 1995) and oxygen exposure times (e.g., Charrie-Duhaut, Lemoine, Adam, Connan, & Albrecht, 2000).
In summary, the ecology and chemistry of mid-Proterozoic marine basins were not all the same. Globally, mid-Proterozoic sedimentary basins reveal a more diverse water column chemistry and preserve a wider range of organic compounds than assumed a decade ago Inevitably, the identification and analysis of a larger sample set of suitable sedimentary successions for biomarkers, microfossils and inorganic geochemistry of different depositional environments, in particular nearshore and oxygenated sedimentary basins, will lead to a greater insight into the relationship between early complex life, biogeochemical cycles and marine and atmospheric redox in the Mesoproterozoic.

| CON CLUS IONS
Indigenous hydrocarbon biomarkers extracted from siltstones in the Velkerri Formation of the McArthur Basin from drillcore Altree 2 have been used to determine the microbial ecology of the ca.  Cox et al., 2016 at similar depths in the drillcore). The 2,3,4-and 2,3,6-trimethyl aryl isoprenoids (TMAI) were absent or in very low concentration in the Velkerri Formation. The low abundance is primary and not caused by thermal destruction.
Instead, TMAI abundances were either affected by degradation of carotenoids during intermittent oxygen exposure at the sedimentwater interface and/or the water column was rarely euxinic in the photic zone and likely only transiently euxinic at depth (>100 m). A comparison of this work with recently published biomarker and trace elemental studies from other mid-Proterozoic basins demonstrates that microbial environments, water column geochemistry and basin redox were heterogeneous.