A community effort towards an improved geological time scale 1 Towards a new geological time scale: A template for improved rock-based subdivision of pre-Cryogenian time

Graham A. Shields1*, Robin A. Strachan2, Susannah M. Porter3, Galen P. Halverson4, Francis A. Macdonald3, Kenneth A. Plumb5, Carlos J. de Alvarenga6, Dhiraj M. Banerjee7, Andrey Bekker8, Alexander Brasier9, Partha P. Chakraborty7, Kent Condie10, Kaushik Das11, Richard Ernst12, Anthony E. Fallick13, Hartwig Frimmel14, Reinhardt Fuck6, Paul F. Hoffman15, Balz S. Kamber16, Anton Kuznetsov17, Ross Mitchell18, Daniel G. Poiré19, Simon W. Poulton20, Robert Riding21, Mukund Sharma22, Craig Storey2, Eva Stueeken23, Rosalie Tostevin24, Elizabeth Turner25, Shuhai Xiao26, Shuanhong Zhang27, Ying Zhou1, Maoyan Zhu28


Introduction
The term 'Precambrian', or more traditionally 'pre-Cambrian' (Glaessner, 1962), is an informal geological term that refers to the time before the beginning of the Cambrian Period at c. 0.54 Ga (Peng et al., 2020). The two pre-Cambrian eonothems (Archean and Proterozoic) have long pedigrees (Sedgwick, 1845;Logan, 1857;Dana, 1872) but were introduced formally only after extensive discussion among members of the Subcommission on Precambrian Stratigraphy (SPS), which was tasked in 1966, with Kalervo Rankama as chair, with standardizing Precambrian nomenclature (Trendall, 1966). James (1978), summarising discussions within the subcommission, outlined five categories of proposal: 1) subdivision by intervals of equal duration (Goldich, 1968; see also Hofmann, 1990;Trendall, 1991); 2) subdivision by major magmatic-tectonic cycles (Stockwell, 1961;1982); 3) subdivision by stratotypes (Dunn et al., 1966; see also Crook, 1989); 4) subdivision by breaks in the geological record defined by radiometric ages (James, 1972); and 5) subdivision based on Earth evolution concepts (Cloud, 1976). One result of those early discussions was that an approximate chronological age of 2500 Ma was assigned to the Archean-Proterozoic boundary (James, 1978). However, further subdivision of the Precambrian in a comparable manner to that achieved for younger rocks, although favoured by some (Hedberg, 1974), proved unworkable (James, 1978) due to 1) the relatively fragmentary nature of the Precambrian rock record, much of which is strongly deformed and metamorphosed, and 2) a scarcity of age-diagnostic fossils. For this reason, a mixed approach was applied: Global Standard Stratigraphic Ages (GSSAs) were introduced to subdivide Precambrian time but the absolute ages of periods were chosen to bracket major magmatic-tectonic episodes (Plumb and James, 1986;Plumb, 1991). Since that decision was ratified, all of pre-Cryogenian Earth history and its geological record has been subdivided using geochronology rather than chronostratigraphy.
The principal Precambrian subdivisions now comprise the informal Hadean and formal Archean and Proterozoic eons (Fig. 1A), which, following the GSSA concept, are defined as units of time rather than stratigraphic packages. The Hadean Eon refers to the interval with no preserved crustal fragments that followed formation of the Earth at c. 4.54 Ga (Patterson, 1956;Manhes et al., 1980). Because the Hadean Eon left no rock record on Earth (other than reworked mineral grains or meteorites), it cannot be regarded as a stratigraphic entity (eonothem) and has never been formally defined or subdivided. It is succeeded by the Archean Eon, which is usually taken to begin at 4.0 Ga and is itself succeeded at precisely 2.5 Ga by the Proterozoic Eon. The Archean Eon is informally divided into four eras (Eoarchean, Paleoarchean, Mesoarchean and Neoarchean), although a three-fold subdivision is widely favoured (van Kranendonk et al., 2012;Strachan et al., 2020). The Proterozoic Eon is currently subdivided into three eras (Paleoproterozoic, Mesoproterozoic and Neoproterozoic) and ten periods (Siderian, Rhyacian, Orosirian, Statherian, Calymmian, Ectasian, Stenian, Tonian, Cryogenian and Ediacaran). The era names were conceived after a proposal from Hans Hofmann in 1987 (Hofmann, 1992;Plumb, 1992), while the period names derive from discussions within the SPS (Plumb 1991). The three Proterozoic eras were originally proposed to begin at 2.5 Ga (Proterozoic I), 1.6 Ga (Proterozoic II) and 0.9 Ga (Proterozoic III), respectively (Plumb and James, 1986). However, the beginning of the Neoproterozoic Era was subsequently moved to 1.0 Ga in the final proposal (Plumb, 1991).
The ages of Precambrian boundaries were selected to delimit major cycles of sedimentation, orogeny and magmatism (Plumb 1991 ; Fig. 1A). However, knowledge has improved considerably over the past thirty years as a result of 1) increasingly precise U-Pb zircon dating, 2) improved isotopic and geochemical proxy records of tectonic, environmental and biological evolution, and 3) new rock and fossil discoveries. As a result, some of these numerical boundaries no longer bracket the events for which they were named. The International Commission on Stratigraphy (ICS) began to address the GSSA immobility problem in 2004 when they ratified the basal Ediacaran GSSP (Global Stratotype Section and Point) on the basis of the stratigraphic expression of a global chemo-oceanographic (and climatic) event in a post-glacial dolostone unit in South Australia (Knoll et al., 2004). Latest geochronology confirms that all typical 'cap dolostone' units were deposited contemporaneously at c. 635.5 Ma (Xiao and Narbonne, GTS 2020). The Ediacaran GSSP is thus one of the most highly resolved system-level divisions of the entire geological record.
The chronostratigraphic (re)definition of the Ediacaran Period (and System) replaced the provisional GSSA (650 Ma) that had been used to mark the end of the Cryogenian Period. This allowed the  to be included within the geological period that owed its name to that glaciation. The 850 Ma age marking the beginning of the Cryogenian was subsequently found to be much older than consensus estimates for the onset of widespread Sturtian 'snowball' glaciation at c. 717 Ma (Macdonald et al., 2010;Halverson et al., 2020) and so it was also removed, following a proposal from the Cryogenian Subcommission (Shields-Zhou et al., 2016). A globally correlative stratigraphic horizon at or beneath this level has not yet been proposed by the Cryogenian Subcommission, although an approximate age of c. 720 Ma for the boundary, pending a ratified GSSP, has been written into the international geological time scale. Despite the lack of a GSSP, the age revision of the Cryogenian Period by 130 million years has been quickly accepted by the geological community worldwide, presumably because it now matches the natural phenomenon for which it was named. International geological bodies have yet to tackle any other issues of non-alignment between names and ages for any other Proterozoic period despite occasional challenges, e.g. Bleeker (2004), van Kranendonk et al. (2012).
In both cases (Cryogenian and Ediacaran GSSPs), and in the case of the earlier ratification of the Precambrian-Cambrian boundary GSSP (Brasier et al., 1994), establishment of a rock-based or chronostratigraphic concept permitted relatively easy consensus around an approximate, stratigraphically calibrated age, before more prolonged and detailed discussions could take place towards eventual GSSP proposal and ratification. In the light of rapidly expanding knowledge about Precambrian Earth history, these three precedents serve to illustrate how the GSSA approach has become outdated and how it is now evident that all Precambrian subdivisions will require revision to a more natural, chronostratigraphic framework (e.g. Bleeker 2004;Van Kranendonk et al., 2012;Ernst et al., 2020). Identified shortcomings with the inflexible GSSA approach are: a) a lack of ties to the rock record and broader Earth history, b) the diachronous nature of the tectonic events on which the current scheme ( Fig. 1A) is based, and c) the lack of any major sedimentological, geochemical and biological criteria that can be used to correlate subdivision boundaries in stratigraphic records. As a result, Proterozoic period nomenclature, although increasingly used in the scientific literature, is commonly out of step with the concepts or phenomena for which they were named, while the underlying basis for both era and period nomenclature is neither universally accepted nor widely understood.
An alternative stratigraphic scheme for the Precambrian was therefore revisited for discussion based on potential Global Boundary Stratotype Sections and Points (GSSPs) (Fig. 1B) (van Kranendonk et al., 2012). Following the rationale of Cloud (1972), the approach taken was to base a revised Precambrian time scale as closely as possible around geobiological events, such as changes to oceans, atmosphere, climate or the carbon cycle that would be near instantaneous compared to changes in geotectonic processes. We agree with the rationale pursued in van Kranendonk et al (2012), which followed an earlier proposal of Bleeker (2004), while noting that some newly proposed subdivisions represent a radical departure from standard practice. This is illustrated by the proposal of a new and exceptionally long 'Rodinian' Period between 1800 Ma and 720 Ma, which replaced five of the preexisting Proterozoic periods. The principle of naming a geological period after a hypothetical supercontinent set a further precedent that is not widely accepted.
Recent progress towards, and widespread acceptance of, chronostratigraphic definitions for two Precambrian periods suggests that the international community can act expeditiously to address the infexibility of the chronometric scheme, while overcoming the confusion generated by the informal erection of new periods and unsupported concepts. This contribution outlines current understanding of Earth history in order to open up for broader consultation new recommendations around 1) a time frame for GSSA removal, 2) rock-based concepts and approximate ages for eon-, era-and possible periodlevel subdivision of pre-Cryogenian time, and 3) new international expert bodies to propose concepts, criteria and candidates for future GSSPs. Our aim is to explore the geological basis behind current chronometric divisions, explore how they might differ in any future rock-based scheme, identify where major issues might arise during the transition, and identify where some immediate changes to the present scheme could be easily updated/formalised, as a framework for future GSSP development. We note that this is not simply a matter of academic interest for geologists. Establishing a robust, coherent and intuitive nomenclature for Earth history is of crucial importance towards improved understanding of our planet's history for schools, universities and academia alike.
Transitioning from a purely chronometric to a chronostratigraphic scheme will inevitably place more emphasis on the rock record and on precise stratigraphic levels within key successions and their global equivalents. In this regard, we accept the arguments made by Zalasiewicz et al. (2004) that units of time and strata are, following the introduction of boundary stratotypes and points or GSSPs, interchangeable and largely superfluous. Specifically, we agree that chronostratigraphic units such as eonothem, erathem or system are confusing and argue that they are particularly inappropriate for subdividing the enormously long time intervals and relatively incomplete rock records of the pre-Cryogenian archive. As a result, we mainly use time subdivisions below (eons, eras, periods), while recalling that their precise definitions and stratigraphically calibrated ages would eventually need to be defined using a level within a global boundary stratotype section.  (van Kranendonk et al., 2012). Golden spike symbols represent ratified (yellow) and potential (pale) GSSPs' Clock symbols represent ratified Proterozoic and recommended Archean GSSAs.

Hadean Eon (> c.4.0 Ga)
Spanning the interval between the formation of the Solar System (c. 4.567 Ga) and the first rock-based evidence for crust on Earth (c. 4.0 Ga), the Hadean Eon was a time of planetary accretion and differentiation. Defining an absolute time scale for the eon is challenging due to the criterion on which it is typically defined: the lack of a preserved rock record on Earth (Cloud, 1976;Strachan et al., 2020). Nevertheless, our understanding of the geological processes that shaped our planet during the Hadean continues to improve. Neither the beginning nor the end of the Hadean are yet formally defined by the International Commission on Stratigraphy (ICS) but the end is usually taken as 4.0 Ga (Fig. 1A). It has been proposed that the Hadean-Archaean boundary be defined as the age of the oldest preserved crustal rock (van Kranendonk et al., 2012). The oldest unambiguously dated rocks are 4.031 ± 0.003 Ga orthogneisses from the Acasta gneiss of the Slave Craton, Canada (Bowring and Williams, 1999), which led van Kranendonk et al. (2012) to propose that the boundary be placed at c. 4.03 Ga (Fig. 1B). However, other authorities have preferred a younger boundary at c. 3.85 Ga to represent the time when crustal preservation on multiple cratonic fragments increased sharply (Bleeker, 2004;Kamber, 2015).
Controversially, older ages of up to ~4.4 Ga have been inferred for mafic gneisses in the Nuvvuagittuq Supracrustal Belt in the Superior Province, Canada (O'Neil et al., 2008(O'Neil et al., , 2012, based on contested interpretations of the short-lived 146 Sm-142 Nd systematics in these highly metamorphosed rocks. Zircon U-Pb ages from felsic orthogneisses in the Nuvvuagittuq Belt suggest a formation age of >3.8 Ga (e.g. Darling et al., 2014), and coupled Lu-Hf and Sm-Nd model ages of ~3.9 Ga for the mafic gneisses have been taken as evidence for an Archean, rather than Hadean origin (e.g. Guitreau et al., 2013). However, O'Neil et al. (2012) report a 147 Sm-143 Nd isochron from intrusive gabbro in the belt of 4.115 ± 0.1 Ga. Taken together, these studies highlight the difficulty in unambiguously dating highly deformed and metamorphosed ancient crustal rocks from very small preserved fragments (Kamber, 2015), while confirming that c. 4.0 Ga (or slightly later) still represents a suitable approximate age for the Hadean-Archean boundary if defined on the basis of potentially correlative rock and mineral assemblages. To debate this boundary further lies, however, beyond the scope and intent of the current paper as the defining concept of the Hadean remains unchallenged.

Archean Eon (c. 4.0 to 2.45/2.5 Ga)
The Archean Eon spans the period of early crustal formation and thickening, leading up to the formation of the first cratons and platform sedimentation. It is characterised by granite-greenstone terranes and the extrusion of ultramafic lavas (komatiites), which are very rare in post-Archean rocks. It is the "end" of granite-greenstone terranes which was traditionally used to define the upper boundary of the Archean at 2.5 Ga. Away from the recognised granite-greenstone terranes or younger platform covers, the Archean is characterised by high-grade, polymetamorphic granite-gneiss complexes, of all ages.
Subdivision of the chronometric Archean Eon was not formalised, pending collection of more data and analysis (Plumb, 1991). Nevertheless, the Subcommission on Precambrian Stratigraphy (SPS) voted subsequently in 1991 and again in 1995 to pursue formal subdivision into four eras: Eoarchean (>3.6 Ga), Paleoarchean (3.6 to 3.2 Ga), Mesoarchean (3.2 to 2.8 Ga), and Neoarchean (2.8 to 2.5 Ga) (Fig. 1A); no reasons for the choice of these boundaries were given. Since these subdivisions have not been formally ratified (Robb et al., 2004), they are considered to be recommendations only (Bleeker, 2004). An alternative subdivision scheme was proposed by van Kranendonk et al. (2012), which simplified these to three: the Paleoarchean (4.03 to 3.49 Ga), Mesoarchean (3.49 to 2.78 Ga), and Neoarchean (2.78 to 2.42 Ga), each composed of a number of periods (Fig. 1B). The base of the Paleoarchean in the 2012 proposal was defined by the age of the oldest extant rocks, the Acasta Gneiss, Canada, while the base of the overlying Mesoarchean was defined at the oldest microbially-influenced textures or structures in stromatolites of the North Pole Dome in western Australia, thus representing the oldest potential "golden spike" (Fig. 1B). The Paleoarchean contained an 'Acastan Period', the lower limit of which was defined by the oldest preserved rocks (Acasta Gneiss, Canada) and an 'Isuan Period', the lower limit of which was defined at 3.81 Ga (the age of the Earth's oldest supracrustal rocks, the Isua Supracrustal Belt in Greenland).
The problem with this approach is that the oldest occurrence of a particular rock type may identify chance preservation rather than any fundamental change in geological process, and older terrains may be identified in the future. It also runs counter to the concept of the international geological time scale as a correlative 'stratigraphic' framework. For the present, we would leave the base of the Archean at ~4.0 Ga, pending formal definition of the Hadean-Archean boundary. This argument would also seem to invalidate the Eoarchean as a fundamental unit of geological time. High grade metamorphic-plutonic complexes, with little internal stratigraphic structure (e.g., Limpopo Belt; Western Gneiss Terrains, Western Australia) become more common toward the top of the interval. A similar problem arises with the placement of the base of the Mesoarchean at c. 3.49 Ga for the North Pole Dome stromatolites; not only the problem of these being the "oldest" stromatolites, but the occurrence of stromatolites is controlled by the particular environment, rather than being a definable moment in evolution. A less ambiguous boundary for the base of the Mesoarchean could be the near-coeval base of the first preserved Barberton and Pilbara granite-greenstone terranes; clearly a significant event in Archean crustal evolution. This is currently considered to be ~3.5 Ga, or perhaps a little older.
Van Kranendonk et al. (2012) also proposed changing the end of the Archean to c. 2.42 Ga, based on the first widespread appearance of 'Huronian' glacial deposits in the rock record (Young, 2019) and the approximately contemporaneous change to an oxygenated atmosphere (Great Oxidation Event or GOE), which followed the end of the world's greatest development of banded iron formation (BIF). This approach seems reasonable based on the rock record, which supports globally significant climatic and atmospheric changes around this time (Gumsley et al., 2017). Although the GOE is sometimes taken to refer to a more prolonged interval of the early Paleoproterozoic, we restrict our usage here to this shorter interval between c. 2.45 to c. 2.32 Ga, with a preference for the earlier age (e.g. Warke et al., 2020b). The GOE has been defined in various ways but in recent years has been presumed to begin when atmospheric oxygen had accumulated sufficiently to prevent the formation and/or preservation of mass independent S-isotope fractionation in sedimentary rocks (Farquhar et al., 2000;Bekker et al., 2004). However, it is currently unclear whether this was a globally synchronous event with estimates for MIF-S disappearance varying from 2.45 to 2.32 Ga due to uncertainties over memory effects caused by oxidative weathering of sulphides and recycling of the MIF signature (Reinhard et al., 2013;Philippot et al., 2018;cf. Torres et al., 2018;Warke et al., 2020b). The MIF-S record has also shown structure within the later Archean that could potentially be used as a future stratigraphic marker (Farquhar et al., 2007;Domagal-Goldman et al., 2008;Halevy et al., 2010;Kurzweil et al., 2013) if the structure is confirmed as a global phenomenon rather than superimposed local signals (see e.g. Gallagher et al., 2017).
While the precise ages and correlation of the GOE and MIF-S events and Paleoproterozoic tillites remain uncertain (Young, 2019), an equally significant event in the transition is the end of the major late Archean BIFs. The Archean-Proterozoic boundary might be best constrained/defined by accurately dated tuffs (~2.45 Ga) at the top of the Hamersley Group BIF (Trendall et al, 2004); BIFs in the Transvaal Basin of South Africa have similar ages. This would imply redefining the Siderian Period and moving it into the Neoarchean (van Kranendonk et al., 2012); a simple solution also to the criticism that 2.5 Ga splits this important rock unit. 2.45 Ga also prevents splitting some younger classic "Archean" terranes which overlap the 2.5 Ga boundary. We note that the peak of BIF deposition would still lie in the Siderian Period should its future chronostratigraphic age match its current chronometric age (2.5 -2.3 Ga). Therefore, we envisage two possibilities with either a terminal Archean Siderian Period ending around 2.45 Ga followed by an Oxygenian Period, or retention of the current names and approximate ages. Whichever scheme wins out will depend upon the precise criteria used to define a chronostratigraphic Archean-Proterozoic boundary, although the current authors favour the former option.
Signal reworking should not affect other geochemical or isotopic markers of upper ocean oxygenation, e.g., redox-sensitive elemental and isotopic enrichments. Geochemical signs of oxidation are recorded intermittently from the Mesoarchean (Ossa Ossa et al., 2019), becoming more expansive during the Neoarchean (Ostrander et al., 2019) and are attributed to transient pulses of oxygen prior to the GOE (the so-called 'whiffs of oxygen') (Anbar et al., 2007;Kendall et al., 2015;Stueeken et al., 2015) or more persistent oxygenation of shallow marine platform environments (Riding et al., 2014). Therefore, the Archean-Proterozoic transition interval is characterised as much by the cessation of anoxic shallow marine environments as by the onset of atmospheric oxygenation. The Neoarchean is also remarkable for the highly negative carbon isotope composition of deposited organic carbon during the interval between c. 2.8 Ga and c. 2.6 Ga (Hayes, 1994) as well as a number of other uniquely extreme stable isotope signatures, possibly related to nascent stages of local environmental oxidation (Thomazo et al., 2009).
Complementary records of Archaean change, which can potentially be divided into three chapters, are provided by geochemical and isotopic studies of magmatic rocks that appear to indicate major secular changes in tectonic processes (Kamber and Tomlinson, 2019). On this basis, Griffin et al. (2014) concurred that the Archean Eon is best divided into: 'Paleoarchean' (4.0-3.6 Ga), 'Mesoarchean' (3.6-3.0 Ga) and 'Neoarchean' (3.0-2.4 Ga) eras. In this interpretation, during the 'Paleoarchean' Era, Earth's dominantly mafic crust acted as a stagnant-to-sluggish lid. Towards the close of the era, zircon grains reveal subtle geochemical signs of a change in tectonic regime interpreted as a move away from granitoid production from oceanic plateaus to some form of arc-like settings (Ranjan et al., 2020). It has been suggested that the subsequent 'Mesoarchean'was dominated by major episodes of mantle overturn and plume activity (van Kranendonk, 2011) that led to development of the subcontinental lithospheric mantle and a steady increase between c. 3.3 and c. 3.0 Ga in the K2O/Na2O ratios of TTG (tonalite-trondhjemite-granodiorite) rock suites (Johnson et al., 2019). However, early evidence of subduction is also interpreted at this time, along with diapiric doming, in adjoining terrains of the Pilbara Craton (Hickman, 2004;van Kranendonk et al., 2004).
The Neoarchean witnessed the onset of some form of plate tectonics, and the development of significant volumes of continental crust, characterised by the first K-rich granitoids (Bédard, 2018). However, gravity-driven doming and plume activity was still an active process in the formation of granite-greenstone terranes ~2720-2600 Ga (Jones et al, 2020). Progressive cratonization is reflected by the development of the first true platform covers -the Witwatersrand and Ventersdorp Supergroups from ~2.95 Ga on the Barberton Craton, the Mount Bruce Supergroup from ~2770 Ga on the Pilbara Craton and the c. 2.93 Ga to c. 2.80 Ga successions of Canada, e.g. the Steep Rock Lake Group (Riding et al., 2014). The Pongola Group (~2980 Ga) probably represents an earlier platform, but is intruded by late-tectonic granites. Emergence of this more buoyant crust may have led to these early carbonate platforms, although pre-3.0 Ga fluvial sediments imply at least some earlier regional emergence (Heubeck and Lowe, 1994). The first development of platform covers on stable cratons provides a logical interim boundary within the Archean and the base of the Neoarchean at ~3.0 Ga.
It seems significant that the youngest widespread granite-greenstone terranes (e.g., Yilgarn,, Kenoran, Bulawayan) are coeval with the basalt-rich Fortescue and Ventersdorp platform covers. This could form a basis for future 3-fold subdivision of the Neoarchean, into three periods based on the rock records of a newly defined Siderian Period (see above), the coeval Fortescue-Ventersdorp groups, and the Witwatersrand-Pongola groups, respectively, although to address period-level Archean subdivision any further lies beyond the scope of the present study.
Clear evidence for the transport of surface crustal material to the deep mantle is found in diamonds within kimberlites. Their stable isotope compositions, geochronology and mineral inclusions imply deep transport to the lower mantle of eclogitized basaltic protoliths that had interacted with seawater as long ago as 3.0 Ga (Shirey and Richardson, 2011;Schulze et al, 2013). Cratonization culminated in a globally stable "supercraton" regime around the Archean-Proterozoic boundary (Bleeker, 2003;Cawood et al., 2018). This configuration set up the lithospheric conditions, due to secular cooling of the mantle, for a switch to a full mode of plate tectonics, which arguably began with rift and drift of these Archean cratons and led eventually to amalgamation after c. 1.9 Ga of Nuna (Columbia), which is widely considered to have been a true supercontinent as opposed to a megacontinent comprising <50% of all continental area.

Proterozoic Eon (2.45/2.5 to c. 0.539 Ga)
The Archean and Proterozoic eons represent fundamentally different Earth systems and the boundary between the two is currently placed at precisely 2.5 Ga. This boundary approximates the change from an early Earth that, although still characteristically Archean (e.g. rare K-rich granitoids but widespread komatiite magmatism, granite-greenstone terranes), had gradually developed many aspects of modern tectonics, to a plate tectonic regime that was characterised by stabilized, emergent continental cratons and supercontinent cycles as well as a notable paucity of the high MgO lavas that are common throughout the Archean. The A-P boundary precedes quite closely but does not correspond precisely to a change in surface environments from a reducing to an oxygenated atmosphere, and from wholly anoxic oceans to a more complex ocean redox structure characterised by oxic, anoxic-ferruginous and anoxic-euxinic portions. Consequently, the boundary must represent a planetary step change that significantly transformed, within 100 to 150 Myr, the Earth's biogeochemical cycles, presumably accompanied by the development of novel microbial pathways and metabolisms, leading eventually to larger and more complex (eukaryotic) forms. The Proterozoic marine sedimentary rock record is also marked by a greater diversity of authigenic minerals (Hazen, 2010;Hazen et al., 2011) and carbonate rock textures (James et al., 1998;Shields, 2002;Hodgskiss et al., 2018). The Proterozoic Eon is divided into three eras, namely the Paleo-, Meso-, and Neo-Proterozoic eras.
i. Paleoproterozoic Era (2.45/2.5 to 1.8/1.6 Ga) The Palaeoproterozoic Era witnessed the transition from an Archean tectonic regime of scattered, small supercratons (Bleeker, 2003) to a more conventional form of plate tectonics, which resulted in formation of Earth's earliest widely accepted supercontinent, Nuna (Hoffman, 1989;Rogers and Santosh, 2002;Zhao et al., 2002;Evans and Mitchell, 2011;Zhang et al., 2012) ), as evidenced by ~2.0-1.8 Ga orogenic belts on all of present-day continents. Redefining the Siderian to contain within in it the BIFs at the top of the Neoarchean requires a new period to be defined and named for the earliest Paleoproterozoic; van Kranendonk et al. (2012) refer to this period as the Oxygenian Period.
The Paleoproterozoic sedimentary record provides clues to significant events, some of which are likely to have been global in scale and can probably be related to the large-scale tectonic processes outlined above. Abundances of molybdenum, uranium, selenium, sulfate and iodate increased in marine sedimentary rocks in multiple Paleoproterozoic basins, indicating increasing ocean reservoirs of those redox-sensitive species (Scott et al., 2008;Partin et al., 2013;Kipp et al., 2017;Hardisty et al., 2017;Blättler et al., 2018). This trend has been interpreted as evidence of oxidative weathering caused by atmospheric oxygenation after the GOE together with expansion of oxic conditions in the marine realm, which stabilised these elements as oxyanions in solution, while titrating redox-sensitive iron, manganese and cerium out of solution (Tsikos et al., 2010;Warke et al., 2020a). Accumulation of iron formations (IFs) peaked around the Archean-Proterozoic boundary, but continued until c. 1.8 Ga (Klein, 2005), after which major BIF deposits are scarce but not entirely absent Canfield et al., 2018). The decline in the abundance of BIF is widely attributed to the titration of ferrous iron from the oceans as pyrite, following the onset of oxidative sulphide mineral weathering after the GOE (Canfield, 1998;Poulton and Canfield, 2011), although hydrothermal sources of iron may have diminished after c. 1.8 Ga (Cawood and Hawkesworth, 2014). The disappearance of redox-sensitive detrital minerals, such as pyrite, uraninite and siderite, has long been attributed to the GOE (Holland, 1984(Holland, , 2006Frimmel, 2005;van Kranendonk et al., 2012), although the onset and duration of the GOE is still inadequately constrained (e.g. Luo et al., 2016) and may not have been synchronous everywhere (Phillipot et al., 2018;Hodgskiss et al., 2019). The onset of the GOE is generally considered to have been approximately contemporaneous with what some have interpreted as the Earth's first global scale glaciations (Bekker and Kaufman, 2007;Brasier et al., 2013;Tang and Chen, 2013;Bekker, 2014;Young, 2019). However, the strongest evidence for a Paleoproterozoic Snowball Earth comes from South Africa where evidence of low-latitude glaciation in the Makganyene Formation (Evans et al., 1997) is found below the first evidence of atmospheric oxygenation (Guo et al., 2009).
Paleoproterozoic glacial episodes were followed by the Earth's largest known positive d 13 C excursion (or excursions, see Martin et al., 2015), the Lomagundi-Jatuli Event (LJE) between c. 2.31-2.22 and c. 2.11-2.06 Ga , which accompanied the first major evaporite sulfate deposits (Melezhik et al., 2005;Schröder et al., 2008;Brasier et al., 2011;Blättler et al., 2018). The LJE is followed by the c. 2.06 Ga Shunga Event that is characterised by a major accumulation of Corgrich sedimentary rocks, the generation of giant petroleum deposits (Melezhik et al. 2004) and pyriterich black shale (Sheen et al., 2019) as well as the first sedimentary phosphorite deposits (Kipp et al., 2020). The c. 2.06 Ga Shunga Event interestingly coincides with the emplacement of two Large Igneous Provinces (LIPs), the Bushveld LIP (Kaapvaal craton) and the Kevitsa LIP on the Karelian craton (e.g. Ernst et al. 2020), which inspired the name of the chronometric Rhyacian Period (2.3 -2.05 Ga) after the Greek word Rhyax (meaning streams of lava) (Plumb, 1991). One potentially distinctive feature of the middle Paleoproterozoic is a disputed tectono-magmatic lull between c. 2.3 to 2.2 Ga (Spencer et al., 2018), during which evidence of continental magmatism and orogenesis is scarce, but not entirely absent (Partin et al., 2014;Moreira et al., 2018). Juvenile magmatism reinitiated after c. 2.2 Ga (Condie et al., 2009;Spencer et al., 2018).
The first macroscopic organic-walled fossils, coiled forms similar to Grypania spiralis, appear within Paleoproterozoic strata by c. 1.89 Ga (Han and Runnegar, 1992;Javaux and Lepot, 2018) to be joined by large, more convincingly eukaryote-grade fossils by the end of the era (Zhu et al., 2016). The Paleoproterozoic fossil record contains the c. 1.88 Ga Gunflint fossil microbes, which are taken to be the oldest unambiguous evidence of iron-oxidising bacteria and oxygenic cyanobacteria (Planavsky et al., 2009;Crosby et al., 2014;Lepot et al., 2017), although cyanobacterial fossils are known also from the c. 2.0 Ga Belcher Group in eastern Hudson Bay (Hofmann, 1975;Hodgskiss et al., 2019).
The end of the Paleoproterozoic Era, as currently defined, is characterised by the widespread formation of unmetamorphosed, shallow-marine sedimentary basins with expansive carbonate platforms over increasingly stable cratons, following amalgamation of Nuna. Some such as the c.1.7c. 1.4 Ga Changcheng-Jixian groups of the Sino-Korean or North China craton, are traditionally considered and mapped as Mesoproterozoic successions (Zhao and Cawood, 2012), despite their slightly earlier origin prior to 1.6 Ga. Many other classic mid-Proterozoic sequences also fall into this category, having begun their depositional history similarly during the chronometric Statherian Period (1.8-1.6 Ga). For example, pre-1.6 Ga units originally envisaged to fall within the chronometric Proterozoic II (Plumb and James, 1986: Figure 1) include the McArthur Basin of Australia (Rawlings, 1999); the lower Riphean Burzyan Group of Russia (Semikhatov et al., 2015); the Espinhaço Supergroup and Araí Group of Brazil and coeval Chela Group of central Africa (Chemale et al., 2012;Guadagnin et al., 2015;Pedreira and de Waele, 2008); the Vindhyan and Cuddupah supergroups of India (Ray, 2006;Chakraborty et al., 2020); and the Uncompahgre Group of SW Colorado, USA, which formed during the late stages of the Yavapai orogeny (1.71-1.68 Ga) (Whitmeyer and Karlstrom, 2007).
The difficulty in assigning a precise age for the Paleoproterozoic -Mesoproterozoic boundary due to the prolonged nature of late Paleoproterozoic orogenies, now known to relate to the amalgamation of Nuna, was recognized early on by the Precambrian Subcommission who expressed "individual preferences … from 1400 Ma to 1800 Ma" (Plumb and James, 1986). In this contribution, we recommend that the end of the Paleoproterozoic be provisionally redefined as ~1.8 Ga, and the Statherian be placed in the Mesoproterozoic, pending future definition of GSSPs. This age is constrained by the latest Orosirian magmatic activity dated at ~1.82 Ga and the base of first platform covers ~1.8 Ga.

ii.
Mesoproterozoic Era (1.8/1.6 to c. 1.0 Ga) The Mesoproterozoic Era represents a period of seeming overall stability in Earth history, during which there were long thought to be few changes in the sedimentary record, biogeochemical cycling, climate and biological evolution (Buick et al., 1995;Brasier and Lindsay, 1998 (Idnurm et al., 1995, Page et al, 2000. Relatively thin sandstone-volcanic sequences were initiated about the same time in Brazil and central Africa, but did not continue up into carbonates (Pedreira et al., 2008). The arenitic Changcheng Group of China began later, after ~ 1.7 Ga, and continued through the carbonate-rich Jixian Group until after ~1.44 Ga (Qu et al., 2014), as did the Lower Riphean of Russia (Semikkatov et al, 2015). The final assembly of Nuna was completed at ~1.5 Ga (Pourteau et al., 2018;Gibson et al., 2020) and followed soon after by thick terrigenous deposition in the Roper (North Australia) and Belt-Purcell (western North America) Basins. The period 1.4-1.25 Ga saw few, short-lived basins worldwide, and was dominated by the breakup of the core of supercontinent Nuna (Evans and Mitchell, 2011) although the breakup may have been relatively incomplete (Ernst et al., 2016;Li et al., 2019) as it was likely followed by re-amalgamation into a different configuration by around 1.0 Ga to form the supercontinent Rodinia . This partial breakup could be reflected in the 1.33-1.30 Ga Derim Derim-Yanliao LIP of northern Australia and the North China craton (Bodorkos et al, 2020).
The final period of the Mesoproterozoic was named Stenian for the "Grenvillian" worldwide linear orogenic belts (Plumb, 1991). The type Grenvillian is specifically defined by two events: Boundary, ~1090 -1020 Ma and Rigolet, ~1000 -980 Ma (Rivers, 2015). The Boundary event was probably the time of collision and is defined by high-grade metamorphic rocks from deep in the crust. The Rigolet boundary was a higher-level event with less movement. An extensional event separated the two, and magmatic emplacement of anothosites, charnockites and metagranites occurred throughout. Earlier events, integral to the belt, extend back to c. 1200 Ma. The Sunsás and Natal-Namaqua belts display similar deep-seated metamorphism and magmatism between c. 1200 -1000 Ma (Cornell et al., 2006;Teixeira et al., 2010). However, in Central Australia, the high-grade 1240-1120 Ma Musgrave Orogeny is followed by the Warkuma LIP. The mafic-ultramafic Giles Complex was emplaced at various levels into the metamorphic rocks, coeval with high-evel p[ost-tectonic granites and the bimodal Bentley Supergroup (Howard et al, 2014).
Increasingly convincing discoveries of fossil eukaryotes, in the form of large, multicellular organic-walled fossil fronds and ornamented acritarchs (Zhu et al., 2016;Miao et al. 2019), first occur in rocks that straddle the chronometric Paleoproterozoic-Mesoproterozoic boundary at 1.6 Ga, and suggest that much of the Mesoproterozoic fossil record remains undiscovered. Current fossil and molecular evidence agree that crown group Archaeplastida (a group that includes the red, green and glaucophyte algae) emerged during the Mesoproterozoic Era (Butterfield, 2000;Eme et al. 2014 (Butterfield et al., 1994;Tang et al., 2020), although see possibly earlier examples of red algae from India (Rafatazmia chitrakootensis and Ramathallus lobatus) at c. 1.6 Ga (Bengtson et al., 2017). Ornamented acritarchs are more common eukaryote-grade fossils and some may prove useful for biostratigraphy, e.g. Tappania plana is a widely reported Mesoproterozoic fossil taxon, which has been found in the Ruyang Group of China (Yin, 1997;Yin et al., 2018), Roper Group of Australia (Javaux et al., 2001;Javaux and Knoll, 2017); Siberia (Nagovitsin, 2009), USA (Adam et al., 2017 and Singhora Group, India (Singh et al., 2019). Therefore, the Mesoproterozoic Era, although often given the epithet 'boring', marks the point in geological time when biostratigraphy begins to seem possible.
As summarised by Cawood and Hawkesworth (2014), the period from 1.7 to 0.75 Ga is characterised by a paucity of passive margins (Bradley, 2008), anoxic-ferruginous and regionally euxinic marine environments, an absence of significant Sr isotope deviations in the seawater record, few highly evolved εHf(t) values in zircon grains, limited orogenic gold and VHMS (but major sedimentary exhalative Pb-Zn) deposits, an absence of glacial deposits and a paucity of massive BIF. Although thought to be Paleoproterozoic in age (Bekker et al., 2003;Papineau, 2010), some of the oldest significant phosphorite deposits may instead have formed around the 1.6 Ga mark in India and Australia (McKenzie et al., 2013;Crosby et al., 2014;Chakraborty et al., 2020;Fareeduddin and Banerjee, 2020). Anorthosites and related intrusive rocks are characteristic of the middle Mesoproterozoic (c. 1.5 -1.2 Ga) and are in places spatially and temporally linked to convergent plate margins (Whitmeyer and Karlstrom, 2007;McLelland et al., 2010;Ashwal and Bybee, 2017). Their development during the Mesoproterozoic was attributed by Cawood and Hawkesworth (2014) to secular cooling of the mantle to a temperature at which continental lithosphere was strong enough to be thickened, but still warm enough to result in melting of the lower thickened crust.
A lack of stratigraphically useful new life forms, major climatic changes, and demonstrably global C isotope excursions throughout the Mesoproterozoic makes it particularly difficult to subdivide. However, the Mesoproterozoic C-isotope record may have been less invariant than has been appreciated until now as exceptions to the apparent montony are emerging (e.g. Zhang, K. et al., 2018;Shang et al., 2019). Regional scale magmatic events such as the c. 1320 Ma and c. 1230 Ma dyke swarms of North China (Zhai et al., 2015;Wang et al., 2016), the Mackenzie dyke swarm in Canada (c. 1.27 Ga) and the approximately contemporaneous Ghanzi-Chobe-Umkondo and Midcontinent Rift Systems (c. 1.12-1.08 Ga) have not been linked to any global-scale isotopic excursions that could be used for correlation, although the coincidence of widespread ca. 1385 Ma LIPs and black shales has been proposed as potential rock-based markers of Mesoproterozoic subdivision, e.g., at the Calymmian-Ectasian boundary .
The final period of the Mesoproterozoic was named 'Stenian' for a series of orogenies (Plumb, 1991) that led eventually to formation of the supercontinent Rodinia, which completed its final amalgamation phase by c. 950 Ma (Li et al., 1999;Evans et al., 2016). The first period of the Neoproterozoic was named 'Tonian', due to the lithospheric stretching now known to be related to the break-up of Rodinia, and originally placed at 0.9 Ga (Plumb and James, 1986). The subdivision of Neoproterozoic time has largely been informed by 1) the occurrence of widespread glacial units now known to be of late Precambrian age (Thomson, 1871;1877;Reusch, 1891;Kulling, 1934;Lee, 1936;Mawson, 1949) and 2) fossils of metazoan affinity that postdate those glaciogenic deposits but predate Cambrian strata (Glaessner, 1962). Harland (1964) first proposed the term "infra-Cambrian" or "Varangian" for the late Precambrian system (Fig. 2) based on two discrete diamictite units, the Smalfjord (Bigganjargga) and Mortensnes, although these are now known to be of late Cryogenian (Marinoan) and Ediacaran age, respectively. Harland proposed that the start of this new period should correspond to the base of the lower of these two glacial horizons on the Varanger Peninsula, NE Norway, first described by Reusch (1891). Dunn et al. (1971) subsequently introduced the terms "Sturtian" and "Marinoan" (named after Sturt Gorge and Marino Rocks near Adelaide) for the two glacial epochs recorded in the late Proterozoic strata of the Adelaide Geosyncline of South Australia, emphasizing their utility as chronostratigraphic markers. Cloud and Glaessner (1982) proposed the term "Ediacarian" for the interval spanning from the upper limit of the last late Precambrian glacial deposit to the base of the Cambrian. This term also originates from South Australia (the Ediacaran Hills) where Ediacara-type fossils were recognized (Sprigg, 1947). Plumb (1991) penned the name "Cryogenian" for the period that included the more widespread of the late Neoproterozoic ice ages and the term "Tonian" (meaning stretching in Greek and in reference to the onset of rifting, now related to the break-up of Rodinia) for the preceding period, setting the chronological boundary between them at precisely 850 Ma. These terms and GSSA boundaries were revised from previously suggested period-rank subdivisions on the geological time scale by the Subcommission on Precambrian Stratigraphy (Plumb and James, 1986).
The number, duration, and intensity of the glaciations have been intensely debated (e.g., Kaufman et al., 1997;Kennedy et al., 1998;Halverson et al., 2005), particularly in the light of the Snowball Earth hypothesis (Hoffman et al., 1998;Hoffman and Schrag, 2002;Fairchild and Kennedy, 2007;Etienne et al., 2007). Notwithstanding these debates, the base of the Ediacaran System (Period) was formally ratified in 2004 in the Adelaide Geosyncline (Knoll et al., 2004; at the same stratigraphic level as originally proposed by Cloud and Glaessner (1982) for their "Ediacarian" period. The terms Cryogenian and Tonian are now widely accepted for the two preceding periods (Shields-Zhou et al., 2012, 2016. Recent proliferation of radioisotopic ages spanning Neoproterozoic glacial deposits has largely resolved the question of the number and timing of late Precambrian glacial intervals. It is now well established that two discrete glaciations of global extent occurred during the Cryogenian Period (i.e., between c. 717 Ma and c. 635 Ma). Despite initial reservations, the international community generally uses the terms Sturtian and Marinoan to refer to these two glacial events ("cryochrons", cf. Hoffman et al., 2017) of the Cryogenian Period. This subdivision, though still informal, appears justifiable in light of the geochronological evidence that 1) the Sturtian glaciation is now thought to have begun at c. 717 Ma (Macdonald et al., 2010(Macdonald et al., , 2018McLennan et al., 2018) and ended at c. 660 Ma Cox et al., 2018;Wang et al., 2019) synchronously worldwide, within the uncertainty of available ages, and that 2) the Marinoan glaciation, though shorterlived and of uncertain duration (between about 4 and 17 Myr; Hoffmann et al., 2004;Condon et al., 2005;Prave et al., 2016;Nelson et al., 2020) also ended synchronously at c. 635.5 Ma (Crockford et al., 2018;Zhou et al., 2019). The start of the Cryogenian Period has now been changed to c. 720 Ma so as to encompass only the glacigenic sequences, pending proposal and ratification of a GSSP.
The preceding Tonian Period now lasts 280 million years. Having originally been envisaged to encapsulate a period of lithospheric thinning (supercontinent break-up), the Tonian now covers the final amalgamation of Rodinia (Evans et al., 2016) and a prolonged interval of relative stability prior to the onset of major break-up after 0.85 Ga, and perhaps as late as 0.75 Ga (Jing et al., 2020).

Fig. 2. Evolution of stratigraphic terminology for the Neoproterozoic Era. Note that age ranges for earlier subdivisions are based on current age estimates. Triangle symbol (Δ) denotes the approximate levels of glaciations relevant to the time scale subdivisions. * Denotes the term "Ediacarian" introduced by Cloud and Glaessner (1982).
A proliferation of sedimentary basins in Rodinia between c. 850 and c. 800 Ma (e.g. the Centralian Superbasin of Australia; the East-Svalbard-East Greenland basin, and the Mackenzie Mountains-Amundsen and associated basins of northern-northwestern Canada, the Nanhua rift basin of South China, and the Central Africa Copperbelt (Lindsay, 2002Rainbird et al., 1996;Wang et al., 2011;Bull et al., 2011;Li et al., 2013) was originally interpreted to record an initial phase of Rodinia break-up (Li et al., 1999;Macdonald et al., 2012), perhaps related to insulation of the underlying mantle (Lindsay, 2002) and/or the influence of a series of similarly aged mantle plumes and associated LIP events that impinged on Rodinia at this time (Li et al., 1999(Li et al., , 2003. The existence of widespread basin-scale evaporite deposits with ages ranging from c. 830 to c. 730 Ma (Prince et al., 2019) is consistent with rifting around this time. However, aside from the western margin of North America (e.g. Macdonald et al., 2012;Timmons et al., 2001;Jefferson et al., 1998) evidence of extension leading to continental separation is lacking and true break-up probably began in earnest only around the start of the Cryogenian (e.g. Merdith et al., 2017), followed by a peak in passive margin abundance at c. 600 Ma (Bradley, 2008). Therefore, Rodinia's tenure as a supercontinent coincided with the Tonian Period, as currently defined, which was named for the tectonic stretching that led to its break-up. Division of the long Tonian into two periods is therefore desirable, although at present many of these c. 850-800 Ma basins lack adequate geochronological control.

Tracers of crustal evolution: cornerstone of the geological time scale
Recent research has focused on understanding episodicity and secular trends in the Precambrian geological record, recognising that the supercontinent cycle and mantle dynamics exert a fundamental control on the evolution of not only the Earth's lithosphere, but also the atmosphere and biosphere, including the spatial distribution of elements and therefore the evolution of mineral deposits (e.g. Huston et al., 2016;Frimmel, 2018), via a series of complex, and incompletely understood, feedbacks (e.g. Worsley et al., 1985;Lindsay and Brasier, 2002;Cawood et al., 2013;Young 2013 Various workers have proposed that the Precambrian can be divided into a number of major intervals based on the dominant tectonic process at any one time. Hawkesworth et al. (2016) suggested five intervals: 1) initial accretion, core/mantle differentiation, development of magma ocean and an undifferentiated mafic crust; 2) plume-dominated tectonics (pre-subduction) at c. 4.5-3.0 Ga; 3) stabilisation of cratons and onset of "hot subduction" between c. 3.0 and c. 1.7 Ga; 4) the "boring billion" at 1.7-0.75 Ga; and 5) Rodinia breakup and development of "cold subduction" from 0.75 Ga onwards. Similarly, van Kranendonk and Kirkland (2016) suggested five intervals, each of which starts with a pulse of mafic-ultramafic magmatism, includes the formation of a supercontinent, and ends with an often protracted period of relative quiescence as the previously formed supercontinent drifts and breaks apart. Following c. 4.03-3.20 Ga -the period from the start of the preserved rock record to the onset of modern-style plate tectonics -these stages are: 1) 3.20-2.82 Ga -the onset of modern-style plate tectonics and the oldest recognised Wilson cycle; 2) 2.82-2.25 Ga -commencing with major crustal growth, emergence of the continents and formation of Superior-type BIFs, and closing with magmatic slowdown and stagnant-lid behaviour; 3) 2.25-1.60 Ga -global mafic/ultramafic magmatism followed by global terrane accretion and the formation of Nuna; 4) 1.60-0.75 Ga -the "Boring Billion" but included partial break-up of Nuna and subsequent formation of Rodinia during the Grenville and other orogenies; 5) 0.75 Ga to present -breakup of Rodinia, the Pangean supercontinent cycle and present transition to Amasia (Safonova and Maruyama 2014; Mitchell et al., 2012). Condie and Aster 2010;Condie, 2014). Black: Total ages (n = 37,830); Green: detrital ancient sedimentary rocks (n = 21,849); Blue: detrital modern sediments (n = 7053; 2x); Red: orogenic granitoids (n = 8928). N (vertical axis) is the number of zircon ages as a function of time for a Gaussian kernel bandwidth of three standard deviations (30 m.y.). Worsley et al. (1985) and Nance et al. (1986) pointed out that processes associated with the supercontinent cycle can be tracked by several isotopic proxies. For example, the distribution of U-Pb zircon ages for the past 4.0 Ga (Fig. 3) in orogenic granitoids and detrital sedimentary rocks record similar peaks at 2.7 Ga (and 2.5 Ga), 1.87 Ga, 1.0 Ga, 0.6 Ga and 0.3 Ga, which correspond to the times of global-scale collisional orogenesis and magmatism associated with the amalgamations of Kenora (Kenorland), Nuna (Columbia), Rodinia, Pannotia/Gondwana and Pangea, respectively. A more recent compilation  interprets these peaks to be pulses of crustal growth and revises the timing of later peaks to 1875 Ma, 1045 Ma, 625 Ma, 265 Ma and 90 Ma. A kernel density estimation analysis (Vermeesch et al., 2016) of almost 600,000 detrital zircons (Spencer et al., 2020) confirms similar peaks at 2.69 Ga, 2.50 Ga, 1.86 Ga, 1.02 Ga, 0.61 Ga, 0.25 Ga and 0.1 Ga, and troughs at 2.27 Ga, 1.55-1.28 Ga, 0.88-0.73 Ga, 0.38 Ga and 0.20 Ga. Variations in the mean initial ɛHf and δ 18 O values from detrital zircon grains in recent sediments show negative troughs and positive peaks, respectively, that correspond to times of supercontinent assembly (Fig. 3). Both proxies are consistent with extensive crustal re-working at the time of assembly with more juvenile contributions representing times of supercontinent breakup and dispersal. Most importantly, all current major subdivisions of geological time, 2.5 Ga, 1.6 Ga, 1.0 Ga, c.539 Ma, c.252 Ma and c.66 Ma, sit within the downslope of troughs that follow peaks in zircon abundance. Note, however, that the time between the 'Nuna' peak at 1.87 Ga and the currently defined Paleoproterozoic-Mesoproterozoic boundary, which precedes a long-lived abundance trough, is anomalously long, and reflects protracted assembly of the Nuna supercontinent.

Fig. 3. Peaks in the distribution (1σ error) of U-Pb zircon ages for orogenic granitoids and detrital zircon grains match intervals of supercontinent assembly, whereas troughs correspond to intervals of supercontinent tenure and break-up (from
The effect of crustal processes on seawater composition is recorded by the 87 Sr/ 86 Sr ratios of marine carbonates. High 87 Sr/ 86 Sr values are attributed to times of increased exhumation of old radiogenic crystalline rocks that would accompany supercontinent amalgamation and disaggregation, while low 87 Sr/ 86 Sr values signify reduced exhumation of old crustal domains that occurs when supercontinent breakup is accompanied by enhanced ocean ridge hydrothermal activity, rift-related magmatism and sea level rise (Veizer, 1989). Although commonly used seawater 87 Sr/ 86 Sr curves (Veizer et al., 1999;Shields and Veizer, 2002;Shields 2007) imply that continental weathering had little influence before the end of the Archean, recent studies (e.g. Satkoski et al., 2016) suggest that continental weathering and low-temperature surface alteration were more important than previously suspected during the Archean. Two prolonged peaks in the Sr isotope composition of seawater correspond with the Paleoproterozoic-Mesoproterozoic and Neoproterozoic-Phanerozoic transitions (Shields, 2007;Kuznetsov et al., 2018). These intervals of enhanced continental weathering of more radiogenic rocks coincide with the amalgamation of Nuna and Pannotia/Gondwana, respectively (e.g. Cawood et al., 2013;Nance and Murphy, 2018). The widespread orogenesis that accompanied amalgamation of Rodinia (Namaqua-Natal (Africa), Grenville-Sveconorwegian (North America-Europe) and Sunsás (South America) does not feature prominently in the seawater Sr isotope curve or zircon εHf compilations, likely because these orogens primarily involved juvenile arcs in external orogens (e.g. Spencer et al, 2013) rather than old radiogenic crustal domains. The dominant influence of lithology over weathering rates on the 87 Sr/ 86 Sr record is consistent with the observed negative covariation between 87 Sr/ 86 Sr and zircon εHf records (Hawkesworth et al., 2016). Supercontinent amalgamation has been proposed to increase nutrient availability (including important bio-limiting nutrients such as iron and phosphorus, e.g. Tyrell, 1999) and organic production via higher rates of chemical weathering, leading to pulses of oxygenation (Campbell and Allen, 2008). However, the concept of greatly increased weathering rates would appear to be in conflict with the silicate weathering feedback, which regulates the long-term carbon cycle (Walker et al., 1981). Moreover, carbon isotope trends, which are commonly used to mirror changes in organic burial and therefore oxygenation (e.g. Des Marais et al., 1992;Lyons et al., 2014), may relate instead to tectonic forcing of carbonate weathering, resulting in an inverse (or no) covariation between d 13 C and carbon burial (Shields and Mills, 2017). The relationship between d 13 C excursions and oxygenation is made more complex when we consider that oxidative weathering of organic carbon could have been limited by low atmospheric oxygen (Bekker and Holland, 2012;Daines et al., 2017). As a consequence, negative as well as positive carbon isotope excursions could indicate Proterozoic oxygenation events, driven instead by pyrite burial, particularly at times of enhanced evaporite sulfate weathering since about 2.0 Ga (Shields et al., 2019). Despite the widespread acceptance of a two-step rise in atmospheric oxygen during the Proterozoic Eon , sophisticated biogeochemical models have seldom been applied to the Precambrian Earth system and few proxies exist to constrain absolute atmospheric oxygen levels .
Geochemical evidence indicates that during much of the Proterozoic Eon, upper oceans were oxygenated due to mixing with the atmosphere, but the deep ocean in contrast still tended towards anoxia with local euxinia (Canfield, 1998;Shen et al., 2002;Arnold et al., 2004;Poulton et al., 2004;Farquhar et al., 2010). However, the extent of surface oxygenation is still highly disputed (Slack et al., 2007;Planavsky et al., 2014;Tang et al., 2016;Zhang et al., 2016;Daines et al., 2017). Deep ocean anoxia suggests that atmospheric oxygen levels were low enough that oxygen became easily exhausted at productive margins where organic remineralisation scavenged first oxygen then sulphate. After the assembly of Rodinia, euxinia became less common, leading to mostly ferruginous conditions (Guilbaud et al., 2015). Under such conditions efficient phosphorus removal may have helped to sustain low atmospheric oxygen levels (Guilbaud et al., 2020). A two-step oxygenation history (first surface oceans and atmosphere and then deep oceans) has also been supported by Fe speciation studies, which show increased amounts of Fe (III) relative to Fe (II) in subaerial volcanic rocks after the GOE, but no change in submarine volcanics until after the NOE (Stolper and Brenhin-Keller, 2018).
How environmental parameters, such as climate and redox conditions, were modulated by the supercontinent cycle remains uncertain but tectonic and environmental changes track the same episodicity, providing optimism that natural subdivision of the Proterozoic Eon would be geologically meaningful. Plume generated LIP magmatism could also help define natural Precambrian (and Phanerozoic) boundaries through their likely effects on the wider environment (Horton, 2015;Ernst and Youbi 2017;Ernst et al. 2020 (Ernst and Youbi 2017). Despite the difficulty matching the isotopic record with LIP emplacement and weathering, the 720 Ma Tonian-Cryogenian boundary, now defined by the start of the Sturtian glaciation, has been linked to the Franklin Large Igneous Province (LIP) of northern Laurentia (Macdonald et al., 2010;Macdonald and Wordsworth, 2017;Ernst and Youbi, 2017) and potentially other LIP fragments (Ernst et al., 2020) either through an increase in planetary albedo associated with the release of aerosols or an increase in global weatherability associated with the tropical emplacement of soluble Ca-and Mg-rich flood basalts. This discovery builds on the recognition that LIPs are coeval with many Phanerozoic chronostratigraphic boundaries and that, although regional in scale, LIPs can have global environmental effects and leave global sedimentary records. Thus, while LIPs are not "golden spikes" in themselves, they can represent proxies for golden spikes in the sedimentary record, which bodes well for Proterozoic stratigraphic correlation along Phanerozoic lines.
There is a clear consensus that the Earth's surface environment evolved in a series of events that were controlled ultimately by magmatism and tectonics. In this, there appears to be no essential difference between major geological transitions of the Phanerozoic and Proterozoic, all of which reflect complex interactions and feedbacks between the lithosphere, atmosphere and biosphere. Despite significant gaps in understanding, various isotopic and geochemical proxies track tectonic (supercontinent) cyclicity and offer prospects for a more robust, rock-based pre-Cryogenian time scale.

Nascent potential of Proterozoic biostratigraphy and chemostratigraphy
Stromatolite textures (and structures) were once thought to be age-diagnostic, although this is now more frequently interpreted to reflect a general tendency towards greater biological control over calcium carbonate precipitation through time (Grotzinger, 1990;Arp et al., 2001;Riding, 2008). Coarse sparry texture (botryoidal fans and, microdigitate stromatolites, dendrites, isopachous laminites, and herringbone calcite) is common in Late Archaean-Mesoproterozoic carbonate systems. Although such textures have commonly been referred to as seafloor cement, they formed at the open sediment-water interface rather than as void-fills, unlike later calcite microspar cements, which peaked in the early Neoproterozoic (James et al., 1998), just as cyanobacterial sheaths began to be calcified, forming clotted, 'thrombolite' mounds of calcimicrobial fabric (Arp et al., 2001). Despite evident trends, few if any sharp temporal divisions can be identified in stromatolite type or microbially induced sedimentary structures (MISS). For example, abiotically controlled seafloor carbonate precipitation did not entirely disappear at the end of the Archean, while the first occurrence of fossilised bacteria (or stromatolites) cannot be used to date the evolutionary origins of cyanobacteria (Blank, 2013;Sánchez-Baracaldo et al., 2017; although see Ward et al., 2016), which from geochemical data ought to predate Neoarchean evidence of oxygenation Riding et al., 2014). Supposedly age-diagnostic sedimentary textures, like stromatolite fabrics, have lost favour, while discoveries of new fossils each year continually resurrect the hope that Precambrian biostratigraphy will be a genuine possibility.
Although Precambrian paleontology is a relatively young field, with only belated acceptance of Ediacaran body fossils (Glaessner, 1962), tremendous advances have been made in recent years. Nevertheless, Precambrian biostratigraphy is still in its infancy with respect to fossil discovery and systematic description. It is not unusual, for example, for a new discovery to extend the stratigraphic range of fossil species by tens or even hundreds of millions of years. Macroscopic forms are too rare or simple to be of stratigraphic use, and while organic-walled microfossils are common in both finegrained siliciclastic sedimentary rocks or concretions, assemblages are typically dominated by simple, smooth-walled vesicles known as Leiosphaeridia, a biologically uninformative acritarch taxon. Nonetheless, with a growing number of taxonomic studies linked to new paleoenvironmental, chemostratigraphic and geochronologic constraints (e.g., Sergeev et al., 2012;Tang et al., , 2015Baludikay et al., 2016;Porter and Riedman 2016;Riedman and Porter, 2016;Beghin et al. 2017a, b;Loron and Moczydłowska, 2017;Loron et al., 2019a;Cohen et al. 2017a,b;Javaux and Knoll, 2017;Miao et al., 2019), a nascent Proterozoic biostratigraphic record is beginning to take shape.

the Cryogenian-Ediacaran boundary) are shown (in open squares) by type of age determination. B. Expanded view of the age constraints that establish the synchronous onset and end of the Cryogenian glaciations.
Molecular clock analyses place the origin of crown-group eukaryotes sometime in the Mesoproterozoic or late Paleoproterozoic (e.g., Berney and Pawlowski, 2006;Parfrey et al., 2011;Eme et al., 2014;Betts et al., 2018). Some phylogenetic data suggest that the first photosynthetic eukaryotes may have emerged in freshwater habitats (Blank, 2013;Sanchez-Baracaldo et al., 2017), which may lower their preservation potential in the rock record. The Stenian-Tonian transition interval is increasingly being viewed as a time of crown group eukaryote diversification (e.g. Cohen and Macdonald, 2015;Butterfield, 2015;Xiao and Tang, 2018). Latest Mesoproterozoic and earliest Neoproterozoic rocks are the first to preserve fossils with clear similarities to particular modern eukaryotic clades, including red and green algae, fungi, amoebozoans, and stramenopiles (Butterfield et al., 1994;Butterfield, 2004;Porter et al., 2003;Nagovitsin, 2009;Loron et al., 2019b;Tang et al., 2020), though the affinities of most early Neoproterozoic fossils remain enigmatic. A number of eukaryotic innovations also appear in the record during this time, including scales, tests, biomineralization, and eukaryovory (Porter and Knoll, 2000;Cohen and Knoll, 2012;Cohen et al., 2017a,b;Porter, 2016). In addition, eukaryote-derived sterane biomarkers appear for the first time c. 810 Ma (Brocks, 2018;Zumberge et al., 2019).
Given these evolutionary changes, it is not surprising that late Mesoproterozoic/early Neoproterozoic fossil assemblages are largely distinct from those of early Mesoproterozoic age (Sergeev et al., 2017), and that several organic-walled microfossils have been proposed as index fossils for this interval. These include the acritarch Trachyhystrichosphaera aimika (spheroidal vesicles with sparse, irregularly distributed, hollow processes), which is found in more than 20 sections worldwide in Stenian and Tonian strata, aged c. 1100 to c. 800 Ma (Butterfield et al., 1994;Riedman and Sadler, 2018) or c. 1150to c. 720 Ma (Pang et al., 2020 and Cerebrosphaera globosa (=C. buickii), robust spheroidal vesicles with distinctive wrinkles, common in late Tonian units c. 800-740 Ma Grey et al., 2011;Riedman and Sadler, 2018) (Fig. 4).
Several other distinctive fossils from rocks c. 780-740 Ma have potential in subdividing the Tonian, but there are too few occurrences known at present to have confidence in their ranges (Riedman and Sadler, 2018). Vorob'eva et al. (2009) noted that many long-ranging early Neoproterozoic and late Mesoproterozoic taxa may be biostratigraphically useful with respect to their last appearances, and in this regard it is worth noting that Riedman and Sadler (2018) found that the disappearance of many ornamented taxa in the later Tonian occurred just before or around the time that the distinctive vaseshaped microfossils appear. The vase-shaped microfossils, constrained to range from c. 790-730 Ma (Riedman and Sadler, 2018;, provide the most promising biostratigraphic marker for subdividing Tonian time, and may be useful in defining the Cryogenian GSSP (Strauss et al., 2014), although it is not clear the extent to which the range is controlled by taphonomic factors.

Chemostratigraphy of the Proterozoic Eon (Eonothem)
With the possible exception of the Neoarchean, chemostratigraphy of the Precambrian is restricted to the Proterozoic Eon. The GOE can be traced using a number of redox-sensitive geochemical proxies, e.g. mass-independent fractionation of sulfur isotopes, trace metal concentrations and isotopes, while related changes to microbial metabolisms and biogeochemical cycles may be apparent in nitrogen, iron and other isotope records.
For later portions of the sedimentary archive, the carbon isotope record has most potential for identifying chemo-oceanographic events during the Proterozoic, beginning with the widespread positive anomaly (anomalies), referred to the Lomagund-Jatuli Event (LJE), which started before c. 2.22 Ga and ended by 2.06 Ga (Karhu and Holland, 1996;Melezhik et al., 2007;Martin et al., 2013). Later negative anomalies have been reported at about 2.0 Ga (Kump et al., 2011;Ouyang et al., 2020), 1.6 Ga (Zhang, K. et al., 2018;Kunzmann et al., 2019), 0.93 Ga (Park et al., 2016). Carbon and strontium isotope stratigraphy (mainly on marine carbonates) are the most widely applied chemostratigraphic tools in the Neoproterozoic, and notwithstanding challenges to their general robustness to diagenetic alteration (e.g., Knauth and Kennedy, 2009;Derry, 2010) and reliability as global seawater proxies (Ahm et al., 2019), consistently reveal similar secular trends at both basinal and global scales.
The shift towards higher sustained δ 13 Ccarb (≥ 5‰) values occurred at c. 850 Ma (though not directly dated), as recorded in the Little Dal Group and equivalent strata of northwestern Canada (Fig. 5;Halverson, 2006;Macdonald et al., 2012;Thomson et al., 2015). However, this trend to high δ 13 Ccarb values is punctuated by a discrete and long-lived interval of near zero to negative δ 13 Ccarb values, referred to as the Bitter Springs Anomaly (BSA) (Halverson et al., 2005) after the Bitter Springs Formation in the Amadeus Basin of central Australia where it was first documented (Hill and Walter, 2000). The BSA is well documented in central Australia, Svalbard, northwestern Canada, Ethiopia  and references therein) and possibly India (George et al., 2018). It is constrained by U-Pb zircon CA-TIMS ages to have initiated after 811.5 Ma (Macdonald et al., 2010) and terminated prior to 788.7 Ma (MacLennan et al., 2018). Using a thermal subsidence-type age model applied to Svalbard, Halverson et al. (2018) estimated the BSA to have begun c. 810 Ma and ended c. 802 Ma, for a duration of 8 million years. Zhou et al., 2020).

Fig. 5. Neoproterozoic seawater strontium isotope curve (blue line) superimposed on known large igneous provinces (red bars) and the carbon isotope record (after
Available strontium isotope data show a plateau near 0.7070 following recovery from the Bitter Springs anomaly that is punctuated by a few minor downturns. A larger decline down to 0.7064-0.7065 broadly coincides with the recovery from the c. 738-735 Ma Russøya negative δ 13 Ccarb excursion MacLennan et al., 2018). Cox et al. (2016) suggested that these late Tonian declines in 87 Sr/ 86 Sr were driven by preferential silicate weathering of continental flood basalt associated with the breakup of Rodinia, with the larger late Tonian drop to 0.7064 corresponding to weathering of juvenile arcs in the Arabian-Nubian Sheild (Park et al., 2019). Furthermore, the early Neoproterozoic 87 Sr/ 86 Sr record is strongly biased by data from a small handful of successions (mainly Svalbard and NW Canada) and suffers from poor age control. Additional high-quality Sr isotope data are necessary to fill in the secular record of Neoproterozoic seawater 87 Sr/ 86 Sr, and have been forthcoming from Russia (Kuznetsov et al., 2017) and the North China craton (Zhou et al., 2020).
The broad contours of the secular trend in Tonian-Cryogenian seawater 87 Sr/ 86 Sr are now well established, dominated by a long-term rise in 87 Sr/ 86 Sr (from ~0.7052 to ~0.7073; Fig. 5). However, strontium isotope chemostratigraphy in the Neoproterozoic is severely limited by the small number of stratigraphic intervals containing limestones that are sufficiently well preserved (i.e., with high Sr/Ca and low Mn/Sr) to record reliably the 87 Sr/ 86 Sr of contemporaneous seawater. Therefore, the record is constructed typically from small numbers of data points from discrete intervals in different successions. When combined with limited age control on most samples, the result is an irregular record with a large number of temporal gaps and limited verification of trends among coeval successions (Fig. 5). Moreover, due to the near absence of syn-glacial carbonate strata, no proxy data for seawater exist for the Cryogenian glacial intervals (i.e. c.717-660 and c.³640-635.5 Ma). Nevertheless, due to the prominent rise in 87 Sr/ 86 Sr through the Neoproterozoic, the strontium isotopic record can potentially distinguish between the early Tonian (i.e., 1000-810 Ma), late Tonian (810-720 Ma) and Cryogenian non-glacial interval (c. 660-650 Ma).

Discussion
The embryonic nature of Proterozoic bio-and chemostratigraphy outlined above illustrates how ratification of pre-Cryogenian GSSPs lies far in the future and beyond the scope of the current review, which is focussed on a template for agreed rock-based criteria to permit the removal of current GSSAs, to be replaced by interim chronostratigraphic units, bounded by approximate ages. Development of a natural Precambrian time scale, especially for periods (systems), is still a 'work in progress' but we consider nevertheless that robust, rock-based subdivision is already possible, desirable and overdue. In working towards this aim, it is however important not to lose sight of the merits of the established chronometric scheme, which has served geologists well over the last 30 years. Indeed, it would appear that most boundaries would change by only small degrees. In order for future units of time (and strata) to be both widely acceptable and scientifically meaningful, they need to be fully defined conceptually, as has been done for the Cryogenian, Ediacaran and Cambrian periods, before they can be pinned down numerically.
As detailed above, the boundary definitions for the Hadean, Archean, Proterozoic and Phanerozoic eons help to broadly delimit four distinct parts of Earth history that are characterised by particular tectonic and biogeochemical regimes. Similarly, the eras of the Proterozoic Eon are recognised to be distinct intervals of tectonic, environmental and biological significance. The goal of any revision to the Precambrian geological time scale should therefore be to minimise disruption to both the current international time scale and existing regional and national stratigraphic norms. In this vein, it may be pertinent to recall the advice given by James (1978), following Trendall (1966), that: (1) "the classification should be the simplest possible that will meet immediate needs [as] every additional complexity provides a basis for disagreement or rejection; (2) The subdivision of time embodied in the classification should reflect major events in Earth's history, yet not be in such a form as to inhibit critical review of that history; (3) The classification must be acceptable to most students of the Precambrian; (4) The nomenclature should not be identified closely with one particular region; and (5) The subdivision scheme should be accompanied by operational criteria, so that assignment to the classification will be guided by objective rather than theoretical considerations". It is in this spirit that we explore below how an improved rock-based geological time scale might depart from the existing chronometric time scale.

Tripartite rock-based subdivision of Archean time
The current geological time scale divides the Archean into four eras, although none of these eras (or the preceding Hadean Eon) has been formally defined or ratified. Successive studies of crustal evolution have tended to favour a simpler three-fold subdivision of the Archean even if the precise numerical boundaries of those eras differ in each study. We support a tripartite subdivision because the rationale for the Eoarchean is conceptually unsound as it seems to be based on the chance preservation of crustal fragments rather than on any fundamental change in Earth history. Although a rock-based onset for the Archean Eon at c. 3.85 Ga is supported by some (Bleeker, 2004;Kamber, 2015), proposing a specific age for the Hadean-Archean boundary lies outside the scope of the current discussion. Moreover it is also too difficult at present to propose a robust subdivision of Archean time based on diagnostic rock assemblages with the possible exception of the Neoarchean Era. We consider that three rock-based Archean eras of equal duration would consititute a robust subdivision to be characterised by the following rock types and geological processes, while acknowledging that future advances in knowledge may amend these preferences in the future: Appearance of granite-greenstone terranes (coeval Pilbara and Barberton cratons with high-grade, polymetamorphic granite-gneiss complexes elsewhere. Widespread uncontroversial stromatolite and other microbial fabrics, largely submerged continents, pulses in TTG magmatism that become progressively more potassium-rich. The base of the Mesoarchean could potentially be chosen for the first appearance of granite-greenstone terranes.

Tripartite rock-based subdivision of Proterozoic time
The current three-fold subdivision of the Proterozoic Eon is broadly accepted with each era following prolonged orogenesis associated with the amalgamation of a supercontinent, marked by zircon abundance peaks at around 2.7-2.5 Ga, 1.90-1.85 Ga and 1.05-1.00 Ga, respectively. The Archean-Proterozoic boundary follows two major zircon abundance peaks associated with the amalgamation of the putative megacontinent Kenora (Kenorland). Rock-based characteristics of the three Proterozoic eras might include: Paleoproterozoic (£2.5 Ga to £1.8 Ga) Plate tectonic regime switch; new platform covers accompanied by onset of widespread (global?) glaciation, oxygenation of the atmosphere (crossing the threshold of disappearance of mass-independent S-isotope fractionation), first widespread sulfate evaporites, carbonates of the Jixian Group (North China) could provide a suitable 'Stratotype', although there are similar carbonate successions in coeval Lower Riphean (Russia). Final suturing of Nuna was followed by thick terrigenous basins; e.g, Roper Group (North Australia), Belt-Purcell Supergroups (North America), Paraguacu-Chapada Diamantina Groups (Brazil). The Statherian-Calymmian boundary is represented by a regional unconformity between the McArthur and Nathan Groups (McArthur Basin) in northern Australia.
Ectasian Period (c.1400 Ma -c. 1200 Ma) after Ectasis (= extension), named for "continued expansion of platform covers". A tectonically quiet period, with little basin development or orogeny; e.g., Xiamaling Group (North China), Yurmatau Grouip (Russia), Kibaran Supergroup (central Africa). Derim Derim-Yanliao and 1230 Ga Maojiagou LIPs. Little to distinguish it from the Calymmian Period, but retained for the present in Figure 6. No suitable "Stratotype" exists for now, although some authors have proposed a level close to the Jixian-Xiamaling group boundary, inferred to be c. 1400 Ma (between 1380 and 1440 Ma).
Stenian Period (c. 1200 Ma -c. 1000 Ma) after Stenos (= narrow), named for "narrow belts of intense metamorphism and deformation", e.g., Grenville, Musgrave, Sunsás, Namaqua orogenies. Conselheiro Mata Group (Brazil) and Keweenawan mafic rift (North America) elsewhere. Defining a chronostratigraphic from high-grade, deep crustal metamorphic rocks is problematic and specifically discouraged by ICS. An interim arbitrary base at c. 1200 Ma is assigned to encompass all the events across the "Grenvillian" belts, and coeval Conselheiro Mata Group.
Neoproterozoic Era (to begin at or after c. 1.0 Ga).
Tonian Period (currently 1000 Ma -c. 720 Ma) after Tonas (= stretch), named for the lithospheric stretching that we now associate with the breaking apart of the supercontinent Rodinia, now thought to have occurred largely after c. 800 Ma based on estimated rift lengths, evaporite deposits and flood basalts (Merdith et al., 2019). Although not currently essential, a new period might conceivably cover the preceding interval of cratonization from the final amalgamation to initial rupture of Rodinia, i.e. approximately £ 1.0 Ga to ³0.8 Ga, with a revised Tonian Period from ³0.8 Ga to c. 0.72 Ga. We tentatively propose the term 'Kratian' in figure 6, following Kratos, the Greek word for strength (cf. Statherian from Statheros).

Concluding remarks and recommendations
1) The four major divisions of Earth history are the Hadean, Archean, Proterozoic and Phanerozoic eons, whereby the Hadean-Archean boundary is taken to represent the start of the terrestrial rock record at c. 4.0 Ga.
2) Two first-order (Archean and Proterozoic eon) and six second-order (Palaeoarchean, Mesoarchean, Neoarchean, Palaeoproterozoic, Mesoproterozoic, Neoproterozoic era) stratigraphic intervals provide intuitive subdivision of post-Hadean / pre-Phanerozoic time. We consider that the Archean Eon would be more parsimoniously subdivided into three informal units of equal duration instead of the current four eras, to be defined further after detailed discussions by a commission of international experts.
3) Major transitions in Earth's tectonic, biological and environmental history occurred at approximately 2.5-2.3, 1.8-1.6 and 1.0-0.8 Ga. We consider, therefore, that current GSSAs at 2.5, 1.6 and 1.0 Ga could be replaced quite expeditiously by rock-based Proterozoic eras beginning at or after c. 2.5 Ga, c. 1.8 Ga and c. 1.0 Ga, respectively, based around these major transitions, all of which occurred following orogenic peaks and during times of waning zircon abundance (postacme, but not yet zenith) in line with Phanerozoic boundaries.

4)
We suggest that current period-level GSSAs be replaced by improved rock-based concepts and interim chronostratigraphic units as soon as practicable, continuing recent progress towards that goal, illustrated, for example, by the establishment of an Ediacaran GSSP in 2004 and chronostratigraphic definition of the base of the Cryogenian at c. 720 Ma in 2012. Although all existing period names could be retained in a future chronostratigraphic scheme, some will need more conceptual underpinning, which would likely result in movement of the Siderian Period into the Archean Eon.

5)
We recommend that a future Paleoproterozoic Era contain only three periods beginning (as now) at or after c. 2.5 Ga, c. 2.3 Ga and c. 2.05 Ga, respectively, so that the era begins near the end of major Archean BIF deposition, the onset of widespread glaciation and the Great Oxidation Event, but ends close to the onset of a prolonged period of cratonic, climatic and isotopic stability and first signs of eukaryotic life in the form of ornamented acritarchs and megascopic fronds. Future attention will likely focus on ensuring that rock-based Paleoproterozoic periods (currently Siderian, Rhyacian and Orosirian) bracket the natural phenomena for which they were named (BIF deposition, magmatism and orogenies, respectively). We propose that the Siderian Period be replaced by the Oxygenian Period (van Kranendonk et al., 2012;Fig. 6b) as the first period of the Paleoproterozoic Era.
6) We recommend that a future Mesoproterozoic Era contain four periods (Statherian at, or more likely after 1.8 Ga, Calymmian at c. 1.6 Ga, Ectasian at c. 1.4 Ga and Stenian at c. 1.2 Ga) so that it begins after major orogenies associated with Nuna amalgamation but before Statherian-age, putative eukaryote-grade fossil assemblages and ends after the Grenville orogeny near the time of the final amalgamation of Rodinia.

7)
We recommend that a future Neoproterozoic Era contain three or four periods (a tentatively named, pre-Tonian Kratian period at or after c. 1.0 Ga, Tonian at c. 0.85 Ga, Cryogenian at c. 0.72 Ga and Ediacaran, which has a ratified GSSP) so that it begins around the final amalgamation of Rodinia and ends traditionally at the Ediacaran-Cambrian boundary.

Fig. 6
Proposed rock-based geologic time scale. Note that era and period boundary ages are approximate ages only and would inevitably change in any internationally agreed chronostratigraphic scheme. Period names in italics represent changes to existing nomenclature. If the first period of the Paleoproterozoic Era were named the 'Oxygenian' (van Kranendonk et al., 2012), we note that the term 'Siderian' would likely be retained for the final period of the Archean Eon. The name 'Kratian' is suggested to precede the Tonian as the first period of the Neoproterozoic Era.