Early Evolution of the Adelaide Superbasin

Abstract

Continental rifts have a significant role in supercontinent breakup and the development of sedimentary basins. The Australian Adelaide Superbasin is one of the largest and best-preserved rift systems that initiated during the breakup of Rodinia, yet substantial challenges still hinder our understanding of its early evolution and place within the Rodinian supercontinent. In the past decade, our understanding of rift and passive margin development, mantle plumes and their role in tectonics, geodynamics of supercontinent breakup, and sequence stratigraphy in tectonic settings has advanced significantly. However, literature on the early evolution of the Adelaide Superbasin has not been updated to reflect these advancements. Using new detrital zircon age data for provenance, combined with existing literature, we examine the earliest tectonic evolution of the Adelaide Superbasin in the context of our modern understanding of rift system development. A new maximum depositional age of 893 ± 9 Ma from the lowermost stratigraphic unit provides a revised limit on the initiation of sedimentation and rifting within the basin. Our model suggests that the basin evolved through an initial pulse of extension exploiting pre-existing crustal weakness to form half-grabens. Tectonic quiescence and stable subsidence followed, with deposition of a sourceward-shifting facies tract. Emplacement and extrusion of the Willouran Large Igneous Province occurred at c. 830 Ma, initiating a new phase of rifting. This rift renewal led to widespread extension and subsidence with the deposition of the Curdimurka Subgroup, which constitutes the main cyclic rift sequence in the Adelaide Superbasin. Our model suggests that the Adelaide Superbasin formed through rift propagation to an apparent triple junction, rather than apical extension outward from this point. In addition, we provide evidence suggesting a late Mesoproterozoic zircon source to the east of the basin, and show that the lowermost stratigraphy of the Centralian Superbasin, which is thought to be deposited coevally, had different primary detrital sources.

1. Introduction

The breakup of the supercontinent Rodinia, and subsequent formation of Gondwana, coincided with critical Earth system changes that led to the Phanerozoic world of extensive macroscopic mineralised life, significantly oxygenated atmosphere and hydrosphere, and a buffered climate devoid of whole-planet glaciations [1,2,3]. Determining any interdependence between these phenomena, e.g., [1,4,5,6], requires constructing full-plate tectonic reconstructions of the globe [7,8], which necessitate a fundamental understanding of the temporal link between tectonically controlled geological features (such as rift basins) and plate tectonic phenomena (such as continental plate sundering and ocean crust formation [8,9]).

The Adelaide Superbasin [10] is one of the largest and best-preserved rift to passive-margin successions to form during the Neoproterozoic breakup of Rodinia, which included large continental rifts between the Australia, Amazonia, Baltica, Kalahari, Laurentia, and Siberia cratons [11,12,13]. The Adelaide Superbasin is thought to have formed the conjugate margin to western Laurentia in Rodinia [8,14,15,16,17,18,19], although other configurations for Rodinia have been suggested, e.g., [13,20,21,22]. Poor chronological control, sparse and ambiguous palaeomagnetic constraints, and a lack of young detrital zircon in the lower units of the Adelaide Superbasin have long hindered the research and testing of these Rodinia reconstructions. Research on the tectonic evolution of the Adelaide Superbasin has seen the geosyncline theory [23] and transition to plate tectonics [24,25], with a few targeted [26,27,28] or more generalized [10,29] studies since.

This research presents new detrital zircon U–Pb and trace element data for the lowermost units of the Adelaide Rift Complex within the Adelaide Superbasin. We use these data, together with existing literature, to provide a refined, early tectonic evolution of the rift system during the deposition of the Callanna Group.

2. Geological Background

2.1. Adelaide Superbasin

The Adelaide Superbasin [10] is a large, Neoproterozoic to middle Cambrian sedimentary system at the southeast margin of Proterozoic Australia which formed as a result of the breakup of the supercontinent Rodinia. The Adelaide Superbasin consists of several named basins and sub-basins that span from the Neoproterozoic to early Cambrian. The largest and oldest of these is the Adelaide Rift Complex, which is contiguous with the relatively undeformed rocks of the Torrens Hinge Zone, Stuart Shelf [23], and Coombalarnie Platform [30]. Two Cambrian basins, the Arrowie Basin and Stansbury Basin, are also considered part of the Adelaide Superbasin [10,31] (Figure 1). Whereas present-day exposure of the sedimentary basin is approximately 600 km from north to south, the basin spans over 1100 km from central Australia to Kangaroo Island. Deposition within the Adelaide Superbasin spans over 300 million years of Earth’s history and stretches from the northernmost regions of South Australia, narrowing in the South Mount Lofty Ranges at the Fleurieu Peninsula and extending onto Kangaroo Island. Further south, links with coeval sequences in Antarctica and eastern Tasmania are unclear, but possible [32]. The Archaean to Mesoproterozoic Gawler Craton lies to the west of the Adelaide Superbasin, and the late Palaeoproterozoic to early Mesoproterozoic Curnamona Province lies to the east. Laurentia is thought to have lain to the east/southeast of the Adelaide Superbasin within Rodinia, and East Antarctica is understood to have been joined to the south of the Gawler Craton as the Mawson Continent (e.g., [10] and references therein). The Adelaide Superbasin began as an intracontinental rift system that successfully progressed to a passive margin basin in its southeast region yet remained a failed rift in the north. Deposition within the basin ceased during the Delamerian orogeny c. 514–490 Ma [25,33,34,35].

The stratigraphy of the Adelaide Superbasin is divided into three supergroups [25], two for the Neoproterozoic sequences and the third for the Cambrian sequences, with numerous group- and subgroup-level divisions. In the Neoproterozoic, the Warrina Supergroup is comprised of the Callanna, Burra, and Poolamacca Groups, and the Heysen Supergroup contains the Umberatana, Wilpena, Torrowangee, and Farnell Groups. Each of these groups are further divided into numerous subgroups, as described by the authors of [10]. The Warrina Supergroup encompasses the Tonian early rift sequences that are largely restricted to fault-bound depositional troughs, and the Heysen Supergroup is comprised of the Cryogenian and Ediacaran glacial, interglacial, and postglacial sedimentary rocks, with a greater area of deposition within a passive margin setting. The timing of rift termination is not well established. However, evidence of large-scale normal faulting is not seen after the early Cryogenian [25]. Here, we focus on the Callanna Group, which is best preserved in the failed arm of the rift system. The reader is referred to Preiss [24], Preiss [25], Counts [36], Lloyd et al. [10], Cowley [37], and references therein for further detail on the geological history of the Adelaide Superbasin.

2.2. Callanna Group

The oldest stratigraphy of the Adelaide Superbasin is represented by the Callanna Group [38], which is further subdivided into the Arkaroola Subgroup [38,39,40] and the Curdimurka Subgroup [38], with the latter inferred to be the younger of the two [41]. For historical reference, the now-outdated “Willouran Series” is equivalent to the Callanna Group, although this has not always been the case [24]. In New South Wales, the Poolamacca Group [42] is thought to be the equivalent of the Arkaroola Subgroup [24]. The known depositional extent of the Callanna Group (Figure 2) is restricted to the eastern (NSW), central, and northern Adelaide Rift Complex (including the Davenport and Denison Ranges), Stuart Shelf, and possibly the eastern Officer Basin.

The Callanna Group is characterised by initially siliciclastic sedimentation transitioning to carbonate and evaporite dominated deposition, with minor, interbedded, mafic to intermediate volcanic and volcanogenic sequences. The Arkaroola Subgroup (Figure 3, Supplementary Figure S1) comprises basal siliciclastic units (e.g., Younghusband Conglomerate, Paralana Quartzite), overlain by a (meta-)carbonate unit (e.g., Wywyana Formation), and finally capped by mafic (meta-)igneous rocks (e.g., Wooltana Volcanics). The basal siliciclastic and middle carbonate sequences are thought to have been deposited in sag basins from the gradual subsidence of a stable craton prior to rifting [25]. Alternately, these sequences may have been deposited as syn-rift sediments penecontemporaneous with faulting [43]. The igneous sequences at the top of the Arkaroola Subgroup are almost exclusively metabasaltic rocks with minor interbedded sediments [24,44]. These igneous sequences are inferred to have been extruded in subaerial settings [44] as continental tholeiitic (flood) basalts [24,25,44,45,46,47]. The Wooltana Volcanics and its equivalent units of the uppermost Arkaroola Subgroup are the most voluminous igneous rocks recognised in the Adelaide Superbasin and have been termed the “Willouran Large Igneous Province (LIP)”, “Willouran Basic Province”, or “Gairdner LIP” [44,45,47,48,49,50,51]. Neoproterozoic mafic volcanics of the Coompana Province c. 860 Ma may also be part of the Willouran LIP [52]. The Willouran LIP (Figure 2) is interpreted to represent the first major phase of rifting within the Adelaide Superbasin, and thus the initiation of Rodinia breakup at the eastern margin of Proterozoic Australia that led to the development of the proto-Pacific Ocean [8]. Presently, the only exposures of complete sections of the Arkaroola Subgroup (Figure 3) are located in the Arkaroola/Mount Painter area, and the Davenport and Denison Ranges (Peake and Denison Inliers) (Supplementary Figure S1). Isolated blocks of the Arkaroola Subgroup are recognised in carbonate megabreccia (diapirs) throughout the Adelaide Superbasin, particularly within the Willouran Ranges. The equivalent Poolamacca Group crops out in the Barrier Ranges of New South Wales (Figure 1).

The Curdimurka Subgroup is thought to overlie the Arkaroola Subgroup and locally exceeds 8 km stratigraphic thickness. As a result of tectonic and salt tectonic dismemberment, no wholly intact section through the Curdimurka Subgroup has been identified [24,25,27,38,54,55,56]. However, composite sections have been developed for the Willouran Ranges [38] (Figure 3), the Davenport and Denison Ranges [54], the Worumba Anticline [56], and the Spalding Inlier [57]. The most intact of these composite sections is within the Willouran Ranges (Supplementary Figure S1). The Curdimurka Subgroup is comprised of a cyclical sequence of evaporitic mixed carbonate and siliciclastic rocks, with minor intermediate to felsic igneous rocks [24,25,27,54,58,59]. The carbonate sequences comprise stromatolitic limestones and dolostones, and cryptalgal dolostone with abundant evaporite mineral pseudomorphs and local tepee structures. The siliciclastic sequences include laminated, pyritic, and carbonaceous siltstone, and sandstones and siltstones with occasional graded bedding, halite casts, and load casts. In addition, feldspathic- and carbonate-cemented cross-bedded sandstone, with occasional heavy mineral laminations and halite casts, are present. The stratigraphic names of the Callanna Group, general geographic locations, and approximate relative stratigraphic positions (correlations) are outlined in Supplementary Figure S1. Supplementary Figure S1 also highlights the significant thickness variations of coeval sequences across the Adelaide Superbasin. Within a given region, significant tectonically controlled local thickness variations occur (e.g., Paralana Quartzite changes thickness by approximately 700 m across the Paralana Fault).

3. Materials and Methods

Three samples were analysed for detrital zircon geochronology, two from the Paralana Quartzite (FR3_007, FR3_008; Figure 2) and one from the Lady Don Quartzite in the eastern part of the superbasin (GSNSWKB001; see Figure 2). The two Paralana Quartzite samples were obtained near the western flank of the Mawson Plateau in the Arkaroola Wilderness Sanctuary, one from the basal member, and one from the top of the same stratigraphic sequence. A fourth volcano-sedimentary sample, 3679330, from the Davenport and Denison Ranges in the far northwest of the superbasin (Figure 2), was also analysed in the hope of obtaining an indication of the crystallisation age of the Cadlareena Volcanics—a presumed equivalent of the Wooltana Volcanics [54]. These samples were selected to investigate the provenance of earliest sedimentary rocks of the Adelaide Superbasin and any spatially related variations in coeval sequences.

Rock samples were first prepared for detrital zircon analysis by crushing the rock samples using a jaw crusher and disk mill. Then, the samples were sieved using nylon mesh of 79 μm and 400 μm. All equipment was thoroughly cleaned by vacuuming, ethanol, and compressed air between each sample. New sieve mesh was used for each sample. Mineral separation was completed by water panning the 79–400 μm fraction and using LST heavy liquid set to a density of 2.85 ± 0.02 g cm⁻³. Zircon was then handpicked and mounted in an epoxy resin. Any grain that remotely resembled a zircon was picked to minimise human bias, an issue highlighted by Sláma and Košler [60] and Dröllner et al. [61]. Where permitted by zircon yields, at least 300 zircons were picked per sample. Otherwise, all zircons in the sample were picked. The mounts were then imaged via cathodoluminescence on either an FEI Quanta 600 scanning electron microscope (for zircon analysed in 2020) or a Cameca SXFive Electron Microprobe (for zircon analysed in 2021). The zircons were then analysed using Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS) to obtain a suite of elemental data for U–Pb geochronology and rare earth element (REE) analysis. All zircons were analysed using a Resonetics M-50 (193 nm ArF excimer) laser ablation system coupled with an Agilent 7900x inductively coupled plasma mass spectrometer. All analytical instruments used are housed at Adelaide Microscopy, University of Adelaide, Australia.

Four standards were used during analysis: GEMOC GJ-1 [62,63], Plešovice [63,64], 91500 [63,65,66], and NIST610 glass [67]. Unknowns were bracketed by two analyses of GJ-1, followed by a combined two to three analyses of Plešovice and 91500, and two analyses of NIST610 for every 20–30 unknowns. GJ-1 was used as the primary calibration standard for U–Pb ratios, and NIST610 was used as the primary calibration standard for Pb isotope ratios and trace element data (See more in Appendix A). Zirconium-91 was used as the internal standard for trace element data with a value of 431,400 ppm (43.14 wt%) ⁹¹Zr assigned to unknowns. Plešovice and 91500 were used as validation standards. A 30 s gas blank, followed by either a 40 s or 30 s ablation (session on 30 March 2021) time, was used with a laser repetition rate of 5 Hz. A spot size of 29 μm and a nominal fluence of 2 J cm⁻² was used for zircon, and a spot size of 43 μm using a nominal fluence of 3.5 J cm⁻² was used for NIST610. Data were processed using LADR [68], version 1.1.06, and output as “Full Analytical Uncertainty”. No common Pb corrections were applied to the data. Reference material ratios for GJ-1, Plešovice, and 91500 were set to the Chemical Abrasion Isotope Dilution Thermal Ionisation Mass Spectrometry (CA-ID-TIMS) values (uncorrected for thorium disequilibria and common-Pb) of Horstwood et al. [63]. Weighted averages and dispersion statistics for all standards are available from the link in “Data Availability”.

Statistical analysis of the zircon U–Pb data followed the method of Lloyd et al. [10]. Data were considered concordant if they were within ± 10%, and a “meaningful” age if the 2σ uncertainty was ≤10%. If a datum satisfied both parameters, it was termed a “Filtered Age”. Maximum depositional ages were determined from a stricter 2% concordance filter, and we used the older age of the three isotope ratios (²⁰⁷Pb/²³⁵U, ²⁰⁶Pb/²³⁸U, ²⁰⁷Pb/²⁰⁶Pb) for a conservative estimate of the youngest single concordant grain. All ages were quoted with 2σ uncertainty. Kernel density estimates (KDEs), and multidimensional scaling plots (MDS) were generated using IsoplotR [69]. Key zircon trace element data are presented graphically using methods following Verdel et al. [70]. In addition, lanthanoid data are represented using violin plots and lambda representation [71,72].

Metadata for the LA-ICP-MS sessions, data for all analyses, cathodoluminescence images, and R code used to generate plots are available from the links in ”Data Availability”.

4. Results

A total of 161 analyses were conducted for sample FR3_008. Of these, 141 analyses passed filtering parameters, with ages ranging from 2914 ± 46 Ma to 892 ± 13 Ma (Figure 4). The primary population peak of this sample was c. 1550 Ma, with a secondary peak c. 1750 Ma, and tertiary peaks c. 1180 Ma and 935 Ma. Four analyses were outside these populations, ranging from 2914 ± 46 Ma to 2237 ± 57 Ma. Notably, a small cluster of zircons occurred c. 900 Ma.

A total of 125 analyses were conducted for sample FR3_007. Of these, 99 analyses passed filtering parameters, with ages ranging from 3090 ± 31 Ma to 1305 ± 17 Ma (Figure 4). The primary population peak of this sample was c. 1680 Ma, with secondary population peaks c. 2480 Ma, 2000 Ma, and 1480 Ma. Three analyses were outside these populations, ranging from 3097 ± 27 Ma to 2819 ± 60 Ma.

A total of 114 analyses were conducted for sample GSNSWKB001. Of these, 85 analyses passed filtering parameters, with ages ranging between 3090 ± 31 Ma and 1302 ± 23 Ma (Figure 4). The primary population peak of this sample was c. 1620 Ma, with a secondary peak c. 1840 Ma. These two peaks formed a bimodal population ranging from 1999 ± 32 Ma to 1302 ± 23 Ma.

From the small quantity of sample that was crushed for sample 3679330, 11 zircons were obtained and analysed, with 10 of these within filtering parameters. The oldest grain yielded an age of 2992 ± 27 Ma, the youngest grain was 1189 ± 18 Ma, and the remainder ranged between 1222 ± 22 Ma and 1725 ± 24 Ma, with a cluster of four grains c. 1680 Ma (Figure 4).

Lanthanoid concentrations are typical for zircons, with a several orders-of-magnitude increase in concentration from light to heavy elements, a slight negative deviation in europium (Eu), and a positive deviation in cerium (Ce) (Figure 5).

5. Discussion

5.1. Provenance and Maximum Depositional Ages

5.1.1. Paralana Quartzite, Including Shanahan Conglomerate Member

Samples FR3_007 and FR3_008 were both sampled from the Paralana Quartzite. However, FR3_008 was sampled from a stratigraphically lower position, mapped as the Shanahan Conglomerate Member. The MDA of the Paralana Quartzite combines the results of both samples.

The youngest zircon in FR3_008 (analysis FR3_008-090, Figure 6) originally yielded ²⁰⁷Pb/²³⁵U, ²⁰⁶Pb/²³⁸U, and ²⁰⁷Pb/²⁰⁶Pb ages of 897 ± 46 Ma, 896 ± 18 Ma, and 889 ± 39 Ma, respectively. To verify the age obtained, this zircon was reanalysed on a subsequent analytical session with two additional analyses. The second analysis (FR3_008_run2-003, Figure 6) yielded ²⁰⁷Pb/²³⁵U, ²⁰⁶Pb/²³⁸U, and ²⁰⁷Pb/²⁰⁶Pb ages of 893 ± 39 Ma, 892 ± 13 Ma, 886 ± 30 Ma, respectively. The third analysis (FR3_008_run2-004, Figure 6) yielded a younger discordant age, likely due to a small inclusion that can be seen in Figure 6.

The two concordant signals had Th/U ratios of ~0.55, and the discordant analysis had a Th/U ratio of ~1.3. A concordia age of 893 ± 9 Ma, MSWD 0.067, p(χ2) 0.98 was calculated from the two concordant analyses, and a traditional uncertainty weighted mean yielded a ²⁰⁶Pb/²³⁸U age of 893 ± 10 Ma, MSWD 0.14 Ma, p(χ2) 0.71. Both calculations propagate external uncertainties [69]. The zircon was euhedral, with simple regular growth zoning presenting a {101} form [75,76]. Although one end of the zircon appeared to have broken off, the aspect ratio was at least 3.3:1. As the concordia age is the statistically “most likely” age [77,78], uses the most amount of available analytical data from the multiple analyses of the single grain, and is in good agreement with individual calculated decay ages and the ²⁰⁶Pb/²³⁸U-weighted mean, it was used as the age of crystallisation and, subsequently, the maximum depositional age for the Paralana Quartzite. This revises the maximum depositional age of the Paralana Quartzite from 1177 ± 28 Ma [10] to 893 ± 9 Ma. There were three additional zircons with ages c. 980–900 Ma, two of which were long, euhedral zircons, and the third was a euhedral overgrowth. This suggests that the youngest zircon, 893 ± 9 Ma, was not a result of contamination.

Both samples, FR3_008 (Shanahan Conglomerate Member) and FR3_007 (Paralana Quartzite), had an overlapping population of zircons c. 1800–1300 Ma, with their primary population peaks centred c. 1580 Ma and c. 1690 Ma, respectively (Figure 4). These primary zircon populations were likely derived locally from the Ninnerie Supersuite and/or Radium Creek Group [79,80,81,82]. The two sample populations differed significantly with the direction of the population tails. Sample FR3_008 tailed toward younger ages, with an additional minor population peak c. 1150 Ma and small cluster of grains c. 900 Ma (Figure 4). There were only four zircons older than c. 1850 Ma present in sample FR3_008. In contrast, sample FR3_007 tailed toward older ages, with an additional minor population peak c. 2500 Ma (Figure 4) and no zircon younger than c. 1300 Ma. Zircons from the older tail of sample FR3_008, particularly the c. 2500 Ma population, were most likely derived from the Gawler Craton (Figure 7), namely the Mulgathing Complex and Sleaford Complex [83,84,85], as previously suggested [10]. The younger c. 1300–1050 Ma zircon population in sample FR3_008 was most likely derived from the Musgrave Province (Figure 7) [86,87,88,89,90]. However, they could have alternately been sourced from a yet-undiscovered but inferred Musgrave-like, late Mesoproterozoic (c. 1300–1000 Ma) source to the east [27,91,92,93]. The five youngest zircons present in sample FR3_008, younger than 1000 Ma, are enigmatic. They have no known local source terrane. Moreover, given the euhedral to subhedral nature of these grains and the breccia-conglomerate nature of the rock, it is unlikely they were transported a great distance. It is possible these zircons were derived from a yet-undiscovered or previously destroyed minor magmatic sequence that would mark initial volcanism of the Adelaide Superbasin that precedes flood basalt emplacement. The zircon populations and lithological differences between the two samples, which were sampled approximately 350 m from each other, suggest a change in the sediment source up stratigraphy to include a greater percentage of more distal source areas, and a loss of the younger source material.

5.1.2. Lady Don Quartzite

Sample GSNSWKB001 was sampled from the Lady Don Quartzite in New South Wales. Based on its lithology and stratigraphic position, it is believed that this formation and the Christine Judith Conglomerate are equivalents to the basal Callanna Group. The maximum depositional age obtained for sample GSNSWKB001 is 1497 ± 52 Ma. There were a few zircons younger than this in the sample, with the youngest being 1302 ± 23 Ma. However, all younger zircons were slightly discordant (>2%, <10%). The sample’s zircon age population is similar to that of the Paralana Quartzite (Figure 4 and Figure 7) samples, with a primary population peak c. 1580 Ma, but includes an additional prominent population peak c. 1850 Ma. There were a few zircons with ages older than 2000 Ma, with one c. 3090 Ma, one c. 2670 Ma, and two c. 2450 Ma. The primary zircon population c. 1580 Ma was likely derived locally from the Ninnerie Supersuite and Radium Creek Group [79,80,81,82], lending support to stratigraphic correlation of the basal Adelaide Superbasin sequences (Figure 7). The additional population c. 1850 Ma was potentially derived from the underlying Willyama Supergroup [94,95], which has been suggested to ultimately be derived from the Arunta Province [96,97]. The few zircon grains older than 2000 Ma were also potentially derived from recycling of the underlying Willyama Supergroup. The rarity of these >2000 Ma zircons suggests that direct transport from the Gawler Craton where these ages were found, namely the Mulgathing Complex, Sleaford Complex, and Cooyerdoo Granite [83,84,85,98], is unlikely.

5.1.3. Cadlareena Volcanics

The small Cadlareena Volcanics sample, 3679330, only yielded 10 zircons that were all interpreted to be inherited/detrital, as there was significant spread with no apparent clustering in the individual ages (Figure 4), and most of the zircons were subhedral and fragmented. The sample’s physical appearance suggests that the rock is a silicified, intermediate volcano-sedimentary rock. Therefore, this result is unsurprising. From this, we interpreted a maximum depositional age of 1189 ± 18 Ma. The ages of the zircon align with those found in the broader region, namely that of the Pitjantjatjara Supersuite of the Musgrave Province [87,99] and the Tunkillia Suite of the Gawler Craton [100,101,102].

5.1.4. Comparison to Basal Central Superbasin Sequences

The Centralian Superbasin developed as an intracontinental basin coeval with the Adelaide Superbasin [103,104], although the superbasins developed relatively independently from each other [25]. Geochronologic control, and thus correlation, of several stratigraphic units across the lower Adelaide Superbasin and Centralian Superbasin remain poor [10,105]. However, the lowermost units are commonly correlated based on stratigraphic similarity and position [24,41,103,105]. In the Centralian Superbasin, these formations are the Heavitree Formation, Dean Quartzite, Vaughn Springs Quartzite, Amesbury Quartzite, Munyu Sandstone, and Kulail Sandstone [105,106,107]. These are thought to be equivalents to the Adelaide Superbasin formations from which the samples analysed in this study were obtained. When the detrital zircon age populations were compared, two main groupings appeared, as the units of the Centralian Superbasin formed one group separate from those of the Adelaide Superbasin (Figure 7). This suggests that the two basins received detritus from differing sources. However, two exceptions occurred: the Heavitree Formation and the Shanahan Conglomerate Member. These two units both plotted (Figure 7) as an intermediary to the more obvious groupings of the Centralian Superbasin and Adelaide Superbasin sequences, suggesting a shared or similar primary detritus source. This is more easily explained for the Heavitree Formation, a relatively mature sandy unit, as the Arunta region, which has somewhat similar zircon age populations as the Gawler Craton, is inferred to be a major source of detritus for the Heavitree Formation [108,109]. This intermediary position in Figure 7 is much harder to reconcile for the Shanahan Conglomerate Member, as this unit is an immature breccia–conglomerate and is unlikely to have received detritus from distal sources. Moreover, as stated earlier, no local source of young zircon is known. This lends some support to the notion of a potential [27,91,92,93] Stenian–Tonian source to the east.

5.2. Zircon Trace Element Geochemistry

Zircon trace element chemistry, particularly of the lanthanoids, uranium (U), thorium (Th), yttrium (Y), oxygen (O), and hafnium (Hf), can be useful in understanding their petrogenesis and provenance, and for crustal evolution [70,110,111,112,113,114,115]. Whereas lanthanoid geochemistry is not thought to be particularly useful in assisting with provenance determinations [116], it is useful at a broader scale for understanding the continental history of a region. Here, we make general observations about the trace element geochemistry of detrital zircon from the lowermost Adelaide Superbasin analysed in this study.

First, as a straightforward measure of continental or oceanic affinity for zircon generation, one can use U/Yb plotted against Y [112,113]. All zircons analysed in this study were inferred to have been generated in the continental crust, as shown by Figure 8. C1 chondrite-normalised [72] concentrations of lanthanoids are typical of zircon (Figure 5) with a positive pattern slope (decreasing λ1 values) from light to heavy lanthanoids, a positive cerium anomaly, and negative europium anomaly [116,117]. Nearly all zircons had a Th/U > 0.07 and were inferred to be originally generated as magmatic rather than metamorphic zircon [118,119]. There was no apparent trend in the lanthanoid pattern slope or curvature (Figure 9), denoted as λ1 (linear slope), λ2 (quadratic slope), and λ3 (cubic slope) [71], with time or sample. Both Eu and Ce anomalies (denoted by Eu* and Ce*) showed a significant spread through time. However, while statistical confidence is limited due to the low number of samples <1000 Ma, these samples generally had low Eu* and Ce* values (“low” is used as in [70], i.e., “strongly negative”). The positive correlation of low Eu* and Ce* values may suggest crystallisation in reduced conditions, thick crust, sediment incorporation, deep mantle plume, effects of fractional crystallisation, and/or competition with plagioclase and/or monazite [70]. The slight increase in Yb/U (Figure 9) in these younger zircons suggests the addition of MORB-like or juvenile mantle-derived magmatism, which is consistent with the type of magmatism accompanying Rodinia rifting.

5.3. Willouran Large Igneous Province and Palaeogeography

Previous authors [20,27,48,121] have advocated for a spatial link of the Willouran Large Igneous Province (LIP) and the Guibei LIP primarily based on igneous geochemistry, palaeomagnetic poles, and geochronology. These authors have advocated for a link between southeast Proterozoic Australia and South China within Rodinia, known as the “missing link” model [20]. Wen and coauthors [21,22] developed an alternative missing link model placing Tarim between Australia and Laurentia instead of South China. However, an increasing number of studies examining detrital zircon, e.g., [122,123,124,125]; geodynamic and kinematic studies, e.g., [8,126,127]; and a recent comprehensive review and update to palaeomagnetic poles [128] have suggested that this position of South China (or Tarim) within the centre of Rodinia is unlikely. Further, this infers that the Willouran LIP and Guibei LIP are not spatially linked as previously suggested [20,27,48,51,121]. The new detrital zircon data in this study further support that the missing link model with Tarim or South China for Rodinia is unlikely. Our detrital zircon data (Figure 4) lack the prominent c. 800 Ma population that is present in samples from Tarim [122] and South China [124,129]. Our data also preserve prominent populations at c. 1580 Ma and c. 1840 Ma that are not prominent within samples from either Tarim or South China.

5.4. Early Evolution of the Adelaide Superbasin

The majority of the Callanna Group has been either eroded, tectonically dismembered, or disrupted by diapirs, and geochronologic controls on deposition remain poor. There are also limited seismic surveys that cross the Adelaide Superbasin [130], and none cross key areas where good stratigraphic control of the Callanna Group is possible (e.g., Willouran Ranges). This makes reconstructing the earliest sequences of the Adelaide Superbasin and its evolution particularly difficult. Here, using existing research on the basin, drawing on literature concerning modern (e.g., East African Rift [131,132,133,134,135,136,137]) and ancient (e.g., Midcontinent Rift [138,139]) rift systems, and new detrital zircon data presented in this paper, we present an updated model for the early evolution of the Adelaide Superbasin.

Initiation of deposition within the Adelaide Superbasin (Figure 10) began between 893 ± 9 Ma and c. 830 Ma. The initial, thin, and geographically restricted, mostly brecciated/conglomeratic clastic sediments (e.g., Shanahan Conglomerate Member) were likely deposited in a series of small, somewhat asymmetric half-grabens with a local detrital source that contained enigmatic young (<1000 Ma) zircons. The half-grabens are thought to have developed by lithospheric thinning under an initial pulse of minor extension focused along pre-existing crustal weaknesses, e.g., Norwest Fault, Isan-Olarian orogen [27,95,140,141,142,143,144,145]. This initial extension was most probably a result of far-field forces [12], although a mantle plume may have played some role through thermal doming [146] or lithospheric weakening. Tectonic quiescence followed, with stable subsidence in the newly created rift, culminating with the deposition of alluvial to fluvial sands and shallow water, and sometimes stromatolitic carbonates (e.g., Paralana Quartzite, Wywyana Formation). This was initially reflected in the change in the zircon spectra of the Paralana Quartzite (Figure 4) to include a greater diversity of detrital sources before shallower water sediments were laid down. It is likely there was transtensional [28,147,148] movement along the Paralana Fault (and its splays) at this time, accounting for the significant thickness variation (~700 m) of the Paralana Quartzite across the fault plane in the Arkaroola area [24]. This interpretation differs from that of Preiss [25] but agrees with Mackay [27] and Job [28], in that we consider the Arkaroola Subgroup to be an early syn-rift, rather than pre-rift, deposition. However, it is worth noting that the amount of extension was minor. The Arkaroola Subgroup is here considered to reflect a sourceward-shifting facies tract (SFT) [53], which fines upward after the initial phase of rift basin development (Figure 3 and Figure 10). Dyke emplacement (Gairdner Dolerite, Amata Dolerite) and extrusion of flood basalts (e.g., Wooltana Volcanics, Beda Basalt) occurs at the top of this first SFT and represents the first major phase of extension in the basin. The flood basalts were extruded in subaerial environments and may have originally formed a continuous sheet [24,25]. After extrusion of the Willouran LIP, rift development continued at an accelerated rate within well-developed grabens, with the deposition of cyclic clastic-carbonate-evaporative sequences of the Curdimurka Subgroup (Figure 3 and Figure 11). This is consistent with detrital zircon and Nd provenance, suggesting a gradual transition from evolved to juvenile, and from broad to restricted detrital sources [10,149] over the long-term evolution of the basin. Evacuation of the magma chambers is thought to be partially responsible for the major graben subsidence [44]. The Curdimurka Subgroup is at least 8 km thick, much thicker than the Arkaroola Subgroup (Supplementary Figure S1), with significant variations across the basin (Figure 11). Magmatism is known to have continued during the deposition of the Curdimurka Subgroup, with bimodal volcanics known from the Willouran Ranges (Rook Tuff), a thin basalt flow in the Spalding Inlier, and xenoclasts of dolerite (thought to belong to the Curdimurka Subgroup) in diapirs/carbonate megabreccia zones [24]. Constraints on the end of Curdimurka Subgroup deposition, and thus the Callanna Group, remain poor. Whereas an exact stratigraphic position has not been determined due to a lack of intact contact relationships [58], the Oodla Wirra Volcanics provide the best determination of a maximum age for the final deposition of the Curdimurka Subgroup where two independent samples yielded ages of 798 ± 5 Ma and 799 ± 4 Ma [58]. This is within uncertainty of the 802 ± 10 Ma age of the Rook Tuff [150] of the lower to mid Curdimurka Subgroup. However, the age determination from the Rook Tuff needs revising as is not reproducible due to the unavailability of the isotopic data from the original analyses. In addition, increased precision and accuracy can be obtained on modern analytical equipment. The minimum age estimate for deposition of the Curdimurka Subgroup is constrained by the Boucaut Volcanics [151] and a porphyry in a basal member of the Skillogalee Dolomite [152] to c. 790 Ma. The exact stratigraphic position of the Boucaut Volcanics remains to be resolved [10]. However, the position of the Skillogalee Dolomite is well constrained. As such, the Callanna Group–Burra Group transition must have occurred between c. 800 Ma and c. 790 Ma and allowed for deposition of the upper Curdimurka Subgroup and entire Emeroo Subgroup. Deposition of the Emeroo Subgroup marks a southward propagation (Figure 12) of the Adelaide Rift Complex following upper Curdimurka Subgroup times. The most southerly deposition of the Curdimurka Subgroup occurred near Spalding [24,25], whereas deposition of the Burra Group occurred as far south as the Adelaide area (Figure 12). A renewed pulse of magmatism (e.g., Boucaut Volcanics, Jarrold Basalt Member, Kooringa Member) occurred at c. 790 Ma [151,152] in the southern and eastern areas of the basin, and likely marked a southern shift in tectonic activity and a period of tectonic quiescence of c. 70–80 million years in the northern Adelaide Superbasin.

In this model, the rift system did not develop as a classic triple junction system through apical extension as suggested by von der Borch [153] and Zhao et al. [154]. Instead, the northern and central areas of the Adelaide Rift Complex initiated as an intra-continental rift that formed along pre-existing crustal weakness and failed to progress to continental breakup, resulting in the present-day aulacogen (Figure 10, Figure 11 and Figure 12). Later development of the Adelaide Rift Complex expanded the extent of the rift system to the south with wider deposition of the Burra Group. This southern region is suggested (Figure 12) to represent the successful rift axis of the Adelaide Superbasin where the proto-Pacific later formed, which is consistent with the kinematic constraints suggested by the authors of [8]. In this model, the triple junctions suggested by von der Borch [153] are a result of the intersection of propagating rifts to form a geometric triple junction.

This model is similar to recent ideas about the development of the Afar triple junction, where the Red Sea meets the Gulf of Aden and the East African Rift system. Traditionally, this area has been viewed as the classic triple junction rift-rift-rift system formed by apical extension away from the triple junction centre [155,156]. However, the geological evidence suggests that at least two of the three arms (Aden Rift, Ethiopian Rift) propagated inward toward the now-seen geometric triple junction, and the chronology of the rift systems does not fit with plume driven apical extension from a central point, e.g., [132,135,136,157]. It appears that the modern Afar triple junction is a geometric place where three rifts, with their predetermined geometries, happened to cross rather than being the point of initiation, e.g., [132,135,136,157], similar to our model for the Adelaide Rift Complex.

6. Conclusions

The development of the Adelaide Superbasin initiated between c. 890–830 Ma with the deposition of the Arkaroola Subgroup in a series of structurally controlled half-grabens in what now constitutes the Adelaide Rift Complex. These structures are likely a manifestation of northeast-southwest (present day)-orientated extensional strain from far-field forces, and potentially also stress from a mantle plume. This phase of extension was limited, and tectonic quiescence followed until extrusion of the Willouran Large Igneous Province (LIP). The Willouran LIP may have been the result of a mantle plume, and its emplacement led to extensive rifting and the subsequent deposition of the Curdimurka Subgroup.

The key findings of this research are:

• Revised constraints on the timing of initial deposition within the Adelaide Superbasin, between ≥893 ± 9 Ma and c. 830 Ma.
• The identification of an enigmatic source of young (<1000 Ma) zircon in the basal stratigraphic unit.
• The Arkaroola Subgroup represents early, syn-rift deposition within half-grabens, developed in an initial pulse of extension that likely exploited pre-existing crustal weakness.
• The central and northern Flinders Ranges formed the initial arm of the rift system but failed to progress to continental breakup.
• Basal Centralian Superbasin and Adelaide Superbasin stratigraphic units had different primary detrital sources.
• Support for a potential late Mesoproterozoic source region to the east of the basin.