Locating the BACE of the Cambrian: Bayan Gol in southwestern Mongolia and global correlation of the Ediacaran – Cambrian boundary

The diversification of animals during the Cambrian Period is one of the most significant evolutionary events in Earth ’ s history. However, the sequence of events leading to the origin of ‘modern ’ ecosystems and the exact temporal relationship between Ediacaran and Cambrian faunas are uncertain, as identification of the Ediacar-an – Cambrian boundary and global correlation through this interval remains problematic. Here we review the controversies surrounding global correlation of the base of the Cambrian and present new high-resolution biostratigraphic, lithostratigraphic and δ 13 C chemostratigraphic data for terminal Ediacaran to basal Cambrian strata in the Zavkhan Basin of Mongolia. This predominantly carbonate sequence, through the Zuun-Arts and Bayangol formations in southwestern Mongolia, captures a distinct, negative δ 13 C excursion close to the top of the Zuun-Arts Formation recognized as the BAsal Cambrian carbon isotope Excursion (BACE). In this location, the nadir of the BACE closely coincides with first occurrence of the characteristic early Cambrian protoconodont Protohertzina anabarica . Despite recent suggestions that there is an evolutionary continuum of biomineralizing animals across the Ediacaran – Cambrian transition, we suggest that this continuum is restricted to tubular forms, and that skeletal taxa such as Protohertzina depict ‘true ’ Cambrian representatives that are restricted entirely to the Cambrian. Employing the first appearance of the trace fossil Treptichnus pedum to define the base of the Cambrian suffers significant drawbacks, particularly in carbonate settings where it is not commonly preserved. As T. pedum is the only proxy available to correlate the Cambrian Global boundary Stratotype Section and Point (GSSP) defined at Fortune Head, Newfoundland, we suggest that the GSSP be redefined elsewhere, in a new stratigraphic section that contains secondary markers that permit global correlation. We propose the nadir of the BACE as the favored candidate to define the base of the Cambrian. However, it is essential that the BACE be complemented with secondary markers. In many global sections the nadir of the BACE and the first occurrence of the genus Protohertzina are closely juxtaposed, as are the BACE and T . pedum . Hence these taxa provide essential biostratigraphic control on the BACE and increase potential for effective global correlation. We also recommend that an Auxiliary boundary Stratotype Section and Point (ASSP) be simultaneously established in order to incorporate additional markers that will aid global correlation of the Ediacaran – Cambrian boundary. The BAY4/ 5 section through the upper Zuun-Arts and Bayangol formations yields key shelly fossils and δ 13 C values and is therefore an

The diversification of animals during the Cambrian Period is one of the most significant evolutionary events in Earth's history.However, the sequence of events leading to the origin of 'modern' ecosystems and the exact temporal relationship between Ediacaran and Cambrian faunas are uncertain, as identification of the Ediacaran-Cambrian boundary and global correlation through this interval remains problematic.Here we review the controversies surrounding global correlation of the base of the Cambrian and present new high-resolution biostratigraphic, lithostratigraphic and δ 13 C chemostratigraphic data for terminal Ediacaran to basal Cambrian strata in the Zavkhan Basin of Mongolia.This predominantly carbonate sequence, through the Zuun-Arts and Bayangol formations in southwestern Mongolia, captures a distinct, negative δ 13 C excursion close to the top of the Zuun-Arts Formation recognized as the BAsal Cambrian carbon isotope Excursion (BACE).In this location, the nadir of the BACE closely coincides with first occurrence of the characteristic early Cambrian protoconodont Protohertzina anabarica.Despite recent suggestions that there is an evolutionary continuum of biomineralizing animals across the Ediacaran-Cambrian transition, we suggest that this continuum is restricted to tubular forms, and that skeletal taxa such as Protohertzina depict 'true' Cambrian representatives that are restricted entirely to the Cambrian.Employing the first appearance of the trace fossil Treptichnus pedum to define the base of the Cambrian suffers significant drawbacks, particularly in carbonate settings where it is not commonly preserved.As T. pedum is the only proxy available to correlate the Cambrian Global boundary Stratotype Section and Point (GSSP) defined at Fortune Head, Newfoundland, we suggest that the GSSP be redefined elsewhere, in a new stratigraphic section that contains secondary markers that permit global correlation.We propose the nadir of the BACE as the favored candidate to define the base of the Cambrian.However, it is essential that the BACE be complemented with secondary markers.In many global sections the nadir of the BACE and the first occurrence of the genus Protohertzina are closely juxtaposed, as are the BACE and T. pedum.Hence these taxa provide essential biostratigraphic control on the BACE and increase potential for effective global correlation.We also recommend that an Auxiliary boundary Stratotype Section and Point (ASSP) be simultaneously established in order to incorporate additional markers that will aid global correlation of the Ediacaran-Cambrian boundary.The BAY4/ 5 section through the upper Zuun-Arts and Bayangol formations yields key shelly fossils and δ 13 C values and is therefore an ideal candidate for consideration as the GSSP for the Ediacaran-Cambrian boundary.

Introduction
The Ediacaran-Cambrian (E-C) interval marks one of the most significant evolutionary events in the history of life on Earth, as the microbe-dominated communities that persisted for over 3 billion years gave way to a Phanerozoic marine world ruled by mobile animals with modern anatomical features (Briggs, 2015).This biological "explosion" of life resulted in the emergence of most of the animal phyla in existence today and their abrupt appearance in the Cambrian has long fascinated and perplexed palaeontologists and evolutionary biologists.The key to determining the timing of crucial evolutionary events across this interval hinges on the development of a chronostratigraphic framework that promotes reliable global correlation of Ediacaran-Cambrian sedimentary successions.This is however, a challenging task, as the dearth of continuous E-C stratigraphic successions, in addition to problems associated with recognizing the E-C boundary continue to hamper our understanding of this key transition.
The base of the Cambrian is defined by the first appearance datum (FAD) of the ichnofossil Treptichnus pedum within the siliciclasticdominated Chapel Island Formation in the Fortune Head section on the Burin Peninsula, Newfoundland (Brasier et al., 1994;Landing, 1994).Since the ratification of this boundary in 1992, the Cambrian GSSP has been the source of constant conjecture and three decades later, the controversy only seems to have amplified (Zhu et al., 2006(Zhu et al., , 2019;;Steiner et al., 2007Steiner et al., , 2020;;Babcock et al., 2014Babcock et al., , 2011;;Peng et al., 2012).The unreliability of T. pedum as a proxy for the base of the Cambrian stems predominantly from its facies dependence and resulting difficulties with correlating clastic settings (where trace fossils are more common) with carbonate sequences (where body fossils of early metazoans occur more frequently and δ 13 C chemostratigraphic signals are obtainable) (e.g., South China, Iran, Kazakhstan and Siberia; Yang et al., 2014aYang et al., , 2016;;Kouchinsky et al., 2017;Steiner et al., 2020;Devaere et al., 2021).δ 13 C chemostratigraphic techniques enable global correlation of regional biostratigraphic schemes, and have been rapidly adopted for the early Cambrian when faunas were often strongly endemic (Maloof et al., 2010a(Maloof et al., , 2010b;;Zhu et al., 2019).However, it is clear that chemostratigraphic data must be employed as part of a multi-proxy dataset in order to extract the most reliable signals for dating and correlation (Babcock and Peng, 2007;Peng et al., 2012;Babcock et al., 2015;Betts et al., 2016Betts et al., , 2017aBetts et al., , 2017bBetts et al., , 2018;;Steiner et al., 2020;Yang and Steiner, 2021).Mixed siliciclastic-carbonate successions have the capacity to provide a diverse range of important chronostratigraphic data, and have therefore become increasingly crucial in the search for and integration of reliable stratigraphic markers for global correlation in the Cambrian.
Southwestern Mongolia is one of only a few areas in the world preserving a thick, continuous mixed siliciclastic and carbonate succession that spans the latest Ediacaran to early Cambrian.As such, it provides the ideal setting to compare the variety of chronostratigraphic tools that have emerged to correlate potential stage boundaries within the Cambrian.Multi-proxy investigations provide the most robust methods of global correlation.However, while all the essential chronostratigraphic elements have been documented in the Zavkhan succession, the decoupling of palaeontological investigations (Voronin et al., 1982;Gibsher and Khomentovsky, 1990;Esakova and Zhegallo, 1996) and chemostratigraphic data (Brasier et al., 1996a;Maloof et al., 2010a;Smith et al., 2016a) has thwarted temporal resolution of these sequences.Furthermore, previous sampling at stratigraphically meaningless "spot localities" or the construction of composite sections that are difficult to replicate has led to discrepancies between age estimates for these rock units and confounded temporal resolution of early animal evolution preserved in one of the most complete Ediacaran-Cambrian successions in the world (Brasier et al., 1996b;Maloof et al., 2010a;Smith et al., 2016a;Landing and Kruse, 2017).To address these issues, we present the first high-resolution multi-proxy chronostratigraphic investigation of the Ediacaran to basal Cambrian Zuun-Arts and Bayangol formations at Bayan Gol in the Zavkhan Basin, southwestern Mongolia (Fig. 1).This study integrates detailed, simultaneously sampled, high-resolution shelly fossil biostratigraphy with δ 13 C chemostratigraphy and lithostratigraphic data.We also review the difficulties surrounding global correlation of the E-C boundary and discuss how the succession in Bayan Gol clarifies problems stemming from these disputes.

Previous work on the Ediacaran-Cambrian deposits of the Zavkhan Basin
The Zavkhan terrane is one of the Proterozoic cratonic fragments in southwestern Mongolia that make up the core of the Central Asian orogenic belt (Macdonald et al., 2009;Bold et al., 2013;Bold et al., 2016aBold et al., , 2016b;;Dorjnamjaa et al., 2016).Sedimentary sequences spanning the E-C transition are widespread across the Zavkhan Basin.The stratigraphic section through the Zuun-Arts and Bayangol formations on the southern fault block in Bayan Gol in particular, has been the focus of many previous studies, laying the foundation for both regional and global correlation (Gibsher et al., 1991;Khomentovsky and Gibsher, 1996;Lindsay et al., 1996;Goldring and Jensen, 1996;Kruse et al., 1996;Brasier et al., 1996b;Esakova and Zhegallo, 1996;Maloof et al., 2010a;Smith et al., 2016a;Fig. 1).Despite extensive previous research, our understanding of this particular section and stratigraphy in the Zavkhan Basin more broadly, continues to be unsettled.This is due to 1) inadequate or at least nonsystematic sampling for skeletal faunas, 2) disagreements surrounding interpretations of the stratigraphy of the sequence and 3) decoupling of biostratigraphic and chemostratigraphic studies.These factors have led to discrepancies between age estimates for the same rock units, hindering regional and global correlation (Voronin et al., 1982;Khomentovsky and Gibsher, 1996;Brasier et al., 1996b;Maloof et al., 2010a;Smith et al., 2016aSmith et al., , 2017;;Landing and Kruse, 2017).

Stratigraphy and mapping
Much of the early work in the Zavkhan Basin was dominated by Soviet and Mongolian groups, who mapped parts of the region and amassed an abundance of geological and palaeontological data from numerous sections across the basin (e.g., Markova et al., 1972;Korobov and Missarzhevsky, 1977;Voronin et al., 1982).Geological formations were subdivided into numbered lithostratigraphic units (Voronin et al., 1982), to assist with the correlation of regional sections and allow the construction of composite biostratigraphic range charts (Voronin et al., 1982;Gibsher and Khomentovsky, 1990;Gibsher et al., 1991).This work primarily focused on providing correlation between the Mongolian and Siberian successions, and by using skeletal fossil assemblages they were able to recognize the Nemakit-Dalydynian, Tommotian, Atdabanian and Botoman stages of the Siberian timescale in the Zavkhan Basin (Voronin et al., 1982;Dorjnamjaa and Bat-Ireedui, 1991).Based on the appearance of distinctive Tommotian fossils, the base of the Tommotian Stage (and hence the base of the Cambrian at that time), was considered by Voronin et al. (1982) to lie at the base of 'unit 5' in Salaany Gol.This unit was later correlated with the base of 'unit 19' in Bayan Gol (Khomentovsky and Gibsher, 1996;Brasier et al., 1996aBrasier et al., , 1996b) ) which is immediately above our measured BAY4/5 section and therefore likely to be ~180 m above the E-C boundary (as defined herein).
These studies during the second half of the 20 th century laid the foundations for future work in the region.However, with time, methods used by these groups have become ambiguous, creating complications when attempting to integrate their results with later investigations.Few details about individual sections were initially provided in these early publications and the sections themselves were frequently not mapped (Markova et al., 1972;Korobov and Missarzhevsky, 1977;Missarzhevsky, 1977).Sampling resolution in these studies is usually insufficient for detailed correlations as locations of palaeontological samples were often designated to broad lithological units (frequently over 50 m in thickness, e.g., Voronin et al., 1982;figs 5, 15) rather than from specific stratigraphic heights along systematically measured sections.Furthermore, samples appear to have been taken sporadically through the successions, heavily concentrated in some units and absent from others (see Voronin et al., 1982, figs 11, 13).In addition, these lithological units were used to develop composite stratigraphic range charts for correlation across the basin regardless of what are likely to be dramatic lateral facies changes, even within a small geographical area (Voronin et al., 1982, fig. 13).
In the early 1990's an International Geoscience Programme (IGCP: Project 303) expedition led by Martin Brasier (Oxford University, UK) and Dorj Dorjnamjaa (Geological Institute, Mongolia) resulted in a Geological Magazine volume dedicated to the Neoproterozoic-Cambrian transition in southwestern Mongolia (Brasier et al., 1996a(Brasier et al., , 1996b;;Evans et al., 1996;Goldring and Jensen, 1996;Khomentovsky and Gibsher, 1996;Kruse et al., 1996;Lindsay et al., 1996).This included the translation of maps and descriptions of key Mongolian sections into English (Khomentovsky and Gibsher, 1996), detailed regional sequence Geological mapping of this part of Bayan Gol has been recently updated (Smith et al., 2016a).Reinvestigation of the area where extensive fault repetition of strata had been documented by Khomentovsky and Gibsher (1996) demonstrated that these strata are more likely to be continuous (Smith et al., 2016a, figs 6B, 8).Faulting would likely have resulted in the distinct, large domical stromatolites of upper unit 18 (the end of the BAY4/5 section in the present study, Bed K of Lindsay et al., 1996) cropping out in their lower unit 18 (Fig. 2; Khomentovsky and Gibsher, 1996, fig. 13), which has not been observed in later studies.Field observations made in the present study included numerous offsets due to small faults, however beds are clearly traceable.Hence, we concur with Smith et al. (2016a fig. 6B) that the lower Bayangol Formation on the west side of the gorge represents continuous stratigraphy (Fig. 2).This reinterpretation dramatically changes the measured thickness of the Bayangol Formation in Bayan Gol (Khomentovsky and Gibsher, 1996 omitted approximately 120 m of strata thought to have represented fault repetition) and further complicates interpretation of the lithostratigraphic framework employed by the authors from the IGCP expedition.

Biostratigraphic studies
One of the main aims of Khomentovsky and Gibsher (1996) was to provide a framework for the stratigraphic succession of skeletal fossils in southwestern Mongolia.The main method for constructing this framework was the identification of key 'marker beds' within numbered lithological units used to correlate disparate packages of strata.The ultimate goal was to construct a single, "super composite" biostratigraphic range chart to which new data and all previously published data could be tied (Khomentovsky and Gibsher, 1996, figs 6, 11).This biostratigraphic scheme included data from the succession exposed in Bayan Gol and was subsequently utilized for the entire Geological Magazine special issue.
In Bayan Gol, the palaeontological sampling of Khomentovsky and Gibsher (1996), fig.6) was restricted to 20 samples through 400 m true thickness of strata, and out of these the approximate stratigraphic levels of only twelve were provided, five of which were from the same horizon.Together with this vague biostratigraphic data from Bayan Gol, RIV RIII 0 m

Unit
No.
Without evidence of repetition, samples RIV-RX were taken much higher in the sequence.

Stromatolitic horizon
Grey limestone

First occurrence of Protohertzina
Strata interpreted as fault repetition and omitted from their lithostratigraphic column We find no evidence of fault repetition.
Outcrop is continuous Khomentovsky and Gibsher's (1996) Khomentovsky and Gibsher (1996) and our interpretation herein.A, Stratigraphic log based on the geological mapping of Khomentovsky and Gibsher, 1996, fig. 6), note the repetition of units 17 and 18 of the Zuun-Arts and Bayangol formations.B, Lithostratigraphic log presented by Khomentovsky and Gibsher, 1996, fig. 13) where the interpreted faulted strata has been omitted resulting in a condensed log and the rearrangement of sample horizons (eg.RIII-X).C, our interpretation that the lower Bayangol Formation represents continuous stratigraphy, resulting in a much thicker sequence and the repositioning of samples RIV-RX.
additional data was incorporated from three other disparate localities; Orolgo Gol, Salaany Gol and Tsagaan Gol (see Khomentovsky and Gibsher, 1996, fig. 13).This biostratigraphic range chart was also expanded upon by Brasier et al. (1996b), fig.6) who included additional fossil occurrences documented by Esakova and Zhegallo (1996) from Bayan Gol, Salaany Gol and Orolgo Gol.Based on the occurrence of eponymous taxa, the Anabarites trisulcatus and Purella zones of the Nemakit-Daldynian Stage of Siberia were identified within the basal Zuun-Arts Formation through to the basal Bayangol Formation, with the base of the Tommotian Stage occurring in their unit 19 (stratigraphically above the measured section presented herein) (Voronin et al., 1982;Khomentovsky and Gibsher, 1996).The absence of Tommotian archaeocyaths however, caused Brasier et al. (1996aBrasier et al. ( , 1996b) ) to question this temporal framework, instead suggesting that most of the Bayangol Formation was Nemakit-Daldynian in age.
It is difficult to assess the accuracy of this composite biostratigraphic column as unit thickness and lithology appear to vary considerably across the basin and key marker beds used for correlation are not entirely clear (Khomentovsky and Gibsher, 1996, fig. 11).These methods were also criticised by Smith et al. (2016a) who observed that extensive lateral facies changes across the basin make lithostratigraphic correlations unreliable.Ambiguous correlations led Smith et al. (2016a) to revise the FAD of a number of shelly fossil groups, most notably the oldest skeletal fauna from the Zavkhan Basin.For example, Anabarites and Cambrotubulus (from the Anabarites trisulcatus Zone; Brasier et al., 1996b) were initially interpreted to have been collected from near the base of the Zuun-Arts Formation in Orolgo Gol (Endonzhamts and Lkhasuren, 1988;Khomentovsky and Gibsher, 1996, figs 11, 13) and were therefore regarded as Precambrian in age (Brasier et al., 1996b).While recent studies have documented cloudinids from the basal Zuun-Arts Formation (Yang et al., 2020), the presence of Anabarites or Cambrotubulus at these levels have not been confirmed (Smith et al., 2016a;Yang et al., 2020 and herein).Because of previously incorrect mapping in Orolgo Gol, Smith et al. (2016a) suggested that these fossil occurrences should be placed 200-300 m higher in the lower Bayangol Formation.In the present study, although poorly preserved tubular fossils occur in the Zuun-Arts Formation (BAY1/104.0-107.8,Figs 3-4), the first definitive occurrence of Anabarites trisulcatus and Cambrotubulus decurvatus (Fig. 5) is also in the lower Bayangol Formation, further supporting recent reinterpretations of these stratigraphic relationships.
Unreliability of previously employed lithostratigraphic correlation methods casts doubt on the utility of a single composite stratigraphic chart such as that presented in Khomentovsky and Gibsher (1996), fig.13) and Brasier et al. (1996b), fig.6).This coupled with erroneous mapping of the Bayan Gol section has resulted in the stratigraphic misplacement of a number of fossil occurrences in their range chart over this crucial interval.Discrepancies between geological mapping and stratigraphic sampling confound correlation (Fig. 2).For example, based on the geological map presented in Khomentovsky and Gibsher, 1996, fig. 6), sample R-VI was collected from 'unit 18' in the lower Bayangol Formation, over one hundred meters northeast of samples R-IV, V, VIII, IX (upper 'unit 18').Therefore, the location of sample R-VI should correspond to strata that are stratigraphically below where R-IV, V, VIII, IX were taken.However, in their super composite column and biostratigraphic chart (Khomentovsky and Gibsher, 1996, fig. 13), sample R-VI is placed stratigraphically higher than samples R-IV, V, VIII, IX, due to the interpreted fault repetition (Fig. 2; Khomentovsky and Gibsher, 1996).The development of this super-composite column further obscures important data: Sample R-VI in Bayan Gol and R-XXI from Tsagaan Gol are represented as precisely the same horizons, despite being collected in localities ~30 km apart (Khomentovsky and Gibsher, 1996, figs 11, 13).Hence, the kinds of fossil taxa that were recovered from each horizon remains unclear.Across both samples (R-VI and XXI) fossil genera reported include Anabarites, Cambrotubulus, Halkieria, Maikhanella and Lophochites; taxa comparable to that documented from the lower Bayangol Formation herein (Fig. 5).However, because the exact stratigraphic location of each sample is uncertain very little meaningful data can be extracted from the assemblages presented in Khomentovsky and Gibsher (1996).The erroneous stratigraphic placement of samples also explains the apparently very early appearance of the mollusc Purella panda (recovered from samples R-IV, V, VIII, IX) in the base of the Bayangol Formation (Khomentovsky and Gibsher, 1996, fig. 13).This horizon corresponds to levels much higher in the section when omitted stratigraphy is included (closer to the top of section BAY4/5 herein).An even earlier occurrence of Purella sp. from the top of the Zuun-Arts Formation is peculiar and requires confirmation, as Khomentovsky and Gibsher (1996) did not image any fossil material.In our study we have not recovered specimens of Purella from the lower Bayangol Formation.All cap-shaped shells in our samples from section BAY4/5 can readily be identified as Maikhanella multa based on the characteristic presence of short sclerite-like scales penetrating the granular matrix of the shell, which differ from the continuous scaly external ornament of Purella (Bengtson, 1992;Devaere et al., 2013).Nevertheless, Khomentovsky and Gibsher (1996) did document a number of key fossil taxa, most notably Protohertzina unguliformis from what appears to be the top of their unit 17 (Fig. 2; Khomentovsky and Gibsher, 1996, sample RIII, fig. 13).Protohertzina unguliformis represents the oldest skeletal fossil occurrence that Khomentovsky and Gibsher (1996) recovered from Bayan Gol (although Zuunia chimidtsereni has recently been documented from near the base of the Zuun-Arts Formation by Yang et al., 2020).Based on the geographic location of this sample (R-III) on the geological map (Khomentovsky and Gibsher, 1996, fig. 6) it should approximately correlate with the first occurrence (FO) of Protohertzina in the present study (Fig. 5).However, confusion generated by erroneous mapping in the region and the lack of systematic sampling has considerably diminished the reliability of these fossil occurrences, including the location of this Protohertzina-bearing sample (Khomentovsky and Gibsher, 1996).The combination of questionable lithostratigraphic correlation techniques used to construct composite range charts, errors with interpreting and mapping geological structures and indiscriminate palaeontological sampling methodologies dramatically reduces the capacity for this biostratigraphic framework to be applied to modern investigations.Consequently, the present study contributes critical new data that addresses these issues.

δ 13 C chemostratigraphic studies
The application of δ 13 C chemostratigraphy for correlation in the Cambrian was still in its infancy when Brasier et al. (1996b) presented the first chemostratigraphic framework for the Zavkhan Basin.Global excursion acronyms (e.g., "BACE") were yet to be coined by Zhu et al. (2006), and so naming conventions for positive and negative peaks differed between studies.In Brasier et al. (1996b) the strong negative δ 13 C excursion in the lower Zuun-Arts Formation in Bayan Gol (-3.3‰) referred to as "anomaly 'W'" was interpreted as representing the major negative excursion seen in the interval between the late Neoproterozoic and earliest Cambrian and was compared to similar curves from Siberia, Iran and Canada (Brasier et al., 1990(Brasier et al., , 1994;;Ripperdan, 1994;Kaufman and Knoll, 1995).Anomaly 'W' and the first appearance of Anabarites trisulcatus provided important anchor points for correlation with the Siberian sections (with excursion 'N' in the Dvortsy section, Braiser et al., 1993;Brasier et al., 1996b, fig. 13).However, it should be noted that the occurrence of Anabarites in the lower Zuun-Arts Formation is most likely erroneous based on unreliable correlations (see above) (Smith et al., 2016a;Yang et al., 2020;herein).Smith et al. (2016a, fig. 8) reported a wide range of isotopic values through the Zuun-Arts Formation in the Zavkhan Basin.In the lower to middle Zuun-Arts Formation at Bayan Gol, values fluctuate between +2‰ and -4‰ (sometimes lower) (Smith et al., 2016a, fig. 8).In contrast, the lower to middle Zuun-Arts Formation in Khunkher Gorge is characterized by relatively constant values that plateau around 0‰ (Smith et al., 2016a, fig. 8).Despite significant variations in isotopic values in the lower-middle Zuun-Arts Formation, a strong negative δ 13 C excursion (minimum value of -7‰) was consistently reported by Smith et al. (2016a) in the upper Zuun-Arts Formation across the Zavkhan Basin.In Bayan Gol the nadir of this excursion occurs 11.3 m below the base of the Bayangol Formation, followed by a series of positive excursions in the lower Bayangol Formation.Smith et al. (2016aSmith et al. ( , 2017, fig. , fig. 12, supplemental material) interpreted this excursion as representing the BACE and correlated it with anomaly 'W' of Brasier et al. (1996b).However, there are discrepancies between the stratigraphic occurrence of this excursion in the Smith et al. (2016a) data and the Brasier et al. (1996b) data.The BACE excursion reported by Smith et al. (2016a) in Bayan Gol occurs ~190 m true thickness above the Boxonia stromatolites that define the base of the Zuun-Arts Formation, whereas anomaly 'W' of Brasier et al. (1996b) occurs only ~43 m above the Boxonia stromatolites (their unit 12) in the same locality.In addition, Brasier et al. (1996b, figs 5, 8) also show the 'W' anomaly in lower-mid levels in the Zuun-Arts Formation in the northern block of Bayan Gol and in Tsagaan Gol (~40 km west of Bayan Gol).In contrast, the BACE in Smith et al. (2016a, fig. 8) is consistently associated with the boundary of the Zuun-Arts and Bayangol formations.Hence, correlation of these excursions is problematic.Brasier et al. (1996b, figs 5, 9,11) document a negative excursion towards the top of the Zuun-Arts Formation that reaches δ 13 C values of -6.2‰ in the Tsagaan Gol section (Brasier et al., 1996b, fig. 5, Table 2) and -3.9‰ in the Bayan Gol section (Brasier et al., 1996b, fig. 9).This excursion was not named, but represents the lower part (before values begin to rise) of the positive 'B' peak that occurs in the lower Bayangol Formation (Brasier et al., 1996b).Based on the magnitude and stratigraphic position of this unnamed negative excursion we consider it a better candidate for correlation with the excursion interpreted as the BACE in Smith et al. (2016a).It is difficult to pinpoint anomaly 'W' of Brasier et al. (1996b) given the wide range of δ 13 C values presented by Smith et al. (2016a) in the lower Zuun-Arts Formation, however it is unlikely to be equivalent to the excursion close to the Zuun-Arts-Bayangol boundary.
Following the BACE, Smith et al. (2016a) show a positive peak (+3.6‰) in the lower Bayangol Formation that correlates with peak 2p in Siberia (maximum values of +2‰; Kouchinsky et al., 2007) and China (maximum values of +4‰; Ishikawa et al., 2008;Li et al., 2013).However, correlating the Siberian 2p peak with data presented by Brasier et al. (1996b) through the lower Bayangol Formation in Bayan Gol is challenging.This area was initially considered to include fault repetition (described above), hence there are considerable uncertainties regarding precise locations of sampling levels.According to Brasier et al.  9), peak 'B' coincided with a distinctive stromatolite horizon ~20 m above the base of the Bayangol Formation.However, recent reinterpretations of these strata indicate that these stromatolites occur ~150 m above the base of the Bayangol Formation (Figs. 1-2, 5).Therefore, it is difficult to determine whether peak 'B' of Brasier et al. (1996b) occurs within their lower unit 18, immediately above the contact with the Zuun-Arts Formation or from their upper unit 18, 150 meters higher up in the section.Unfortunately, little can be gleaned from nearby stratigraphic sections; the levels including excursion 'B' are either faulted (northern block of Bayan Gol [Brasier et al., 1996b,   or marred by sampling issues (Khevtte-Tsakhir-Nuruu ridge), resulting in a unique isotopic signal, dissimilar to other sections (Brasier et al., 1996b, fig. 10).The chemostratigraphic signal in Tsagaan Gol is perhaps the most continuous and reliable reported by Brasier et al., 1996b, fig. 5) and does show two positive peaks ('B' and 'C') through the lower Bayangol Formation separated by a broad negative excursion that reaches values of -3‰.Hence, peak 'B' is most likely to correlate with peak 2p based on isotopic values and stratigraphic position in the lower Bayangol Formation.

Sampling
Preliminary fieldwork in Bayan Gol, Zavkhan Basin, southwestern Mongolia was conducted in August, 2017 and comprehensive measurement and collection through Ediacaran-Cambrian strata took place in August 2018.Two sections were measured.The BAY1 section captures the lowermost 155.8 m of the Zuun-Arts Formation (Fig. 3) and the BAY4/5 section herein captures the uppermost 11.96 m of the Zuun-Arts Formation and 149.59 m of the overlying Bayangol Formation (Fig. 5).BAY1 was measured on the northeast side of Bayan Gol (base of section coordinates are N46 • 42'41.1"/E96• 18'57.6";Fig. 1), commencing at the karstic horizon that defines the boundary between the Shuurgat Formation and the distinctive Boxonia stromatolites at the base of the Zuun-Arts Formation (Smith et al., 2016a;Adachi et al., 2019;Yang et al., 2020).Initial exploration of the lower Zuun-Arts Formation revealed relatively continuous strata.However, strong faulting further upsection disrupted strata, making individual beds increasingly difficult to follow.Hence, BAY1 was terminated at BAY1/226.1 (155.8 m true thickness from base) (Figs 1, 3).
BAY4/5 was measured on the western side of Bayan Gol (Fig. 1) and intersects the uppermost 11.96 m of the Zuun-Arts Formation (Fig. 5).Obvious faults stratigraphically below the BAY4/5 section (Fig. 1) prevented the section being extended further down into the underlying Zuun-Arts Formation.It is most likely that the two sections (BAY1 and BAY4/5) do not stratigraphically overlap.We attempted to replicate previous work that managed to develop stratigraphically thicker "composite columns" for the Zuun-Arts Formation in Bayan Gol by correlating separate stratigraphic sections measured in adjacent fault blocks (Brasier et al., 1996b;Smith et al., 2016a;Fig. 1).However, our investigation of the area showed the uppermost Zuun-Arts Formation to be characterised by a series of massive, blocky, carbonate parasequences with few distinctive features.Distinctive "marker horizons" are rare or absent, precluding confident replication of composite columns created via correlation across faults.Samples taken through the lower Zuun-Arts Formation were processed for both stable isotopes and shelly fossil extraction and have yielded only a few poorly preserved tubular fossils (Fig. 4).
In the present study our aim was to reduce as much as possible the potential for errors made when correlating across faulted blocks through lithologically similar units.This results in stratigraphically reduced, but more reliable and reproducible measured sections.Locations where sections have been measured were carefully chosen for their capacity to provide continuous strata in terrain that could be measured with a tape, and where the need for offsets were minimal.If offsets were required, they were made on strongly outcropping beds, or by following clear lithological signals along strike to ensure that no levels were missed or duplicated.Hence, BAY4/5 (Fig. 1) was measured on a single fault block through a relatively undisturbed, continuous sequence ensuring that biostratigraphic, chemostratigraphic and lithostratigraphic data presented herein may be reproduced by future workers.
BAY 1 and BAY4/5 were measured with a tape and detailed sedimentary logs were recorded including frequent measurement of dip angles in order to calculate true thicknesses.Samples were collected at 5-10 m intervals from logged horizons.At each sampled horizon material was collected for acquisition of multi-proxy data, i.e. limestone blocks for SSF extraction, a fresh rock sample for bulk stable isotope (δ 13 C and δ 18 O) analysis and an oriented sample for thin section production.In total, 76 horizons were sampled over 161.55 m stratigraphic thickness in BAY4/5 (Fig. 5) and 21 horizons were sampled (17 for stable isotope analysis and SSF extraction plus an extra 4 for SSF extraction only) in BAY1 over 155.8 m true thickness (Fig. 3).Macroscopic algae and trace fossils were also collected and their stratigraphic occurrences logged onto the section.
Macrofossil samples were imaged with a Canon EOS 5D Mark III DSLR camera at the Swedish Museum of Natural History in Stockholm (NRM) and with a Canon EOS 7D Mark II camera in the Institute of Palaeontology of the Mongolian Academy of Sciences in Ulaanbaatar.Samples for shelly fossil extraction were digested in 5% acetic acid at the NRM following acetic maceration methodology of Jeppsson et al. (1999) and hand-picked fossils from the resulting insoluble residues were gold-coated and imaged at the NRM using a FEI Quanta FEG Scanning Electron Microscope at an accelerating voltage of 15 kV.Uncoverslipped thin sections of samples from selected intervals were made at the University of New England in Armidale, Australia (UNE) and imaged with a Nikon 90i microscope controlled via NIS Elements software.Elemental maps of polished thin sections were produced with the Bruker M4 Tornado at Macquarie University.All specimens are stored at the Swedish Museum of Natural History (NRM).

Stable isotope methods
Samples for δ 13 C and δ 18 O isotope chemostratigraphy were collected simultaneously with shelly fossil and lithological samples through the Zuun-Arts and Bayangol formations.A total of 76 isotope samples were taken over 161.55 m stratigraphic thickness in BAY4/5 and 17 samples in BAY1 (Figs 3,5; see Supplementary Information).Stable isotope analyses (carbon [δ 13 Ccarb] and oxygen [δ 18 Ocarb]) for BAY4/5 were conducted at the Nanjing Institute of Palaeontology and Stratigraphy (NIGPAS) and for BAY1 at the University of Copenhagen, Denmark.Fresh surfaces of rock samples were carefully assessed prior to microdrilling to avoid weathering rinds, veins and other secondary textures.1-2 g of powder from each sample was collected with a dental drill and each sample was analysed simultaneously for δ 13 C and δ 18 O data with a Thermo Scientific MAT-253 mass spectrometer with Kiel IV Carbonate Device.For each sample, an aliquot of 80 to 100 μg of powder was reacted with orthophosphoric acid for 150-200 s at 72 • C. The CO generated from this reaction gas was analyzed for δ 13 C carb and δ 18 O carb using a MAT-253 mass spectrometer.The VPDB standard (GBW-04405, with a δ 13 C carb value of 0.57 ± 0.03‰ and a δ 18 O carb value of − 8.49 ± 0.13‰) was used for analytical calibration, with an analytical precision better than ±0.08‰ for δ 13 C carb and ±0.1‰ for δ 18 O carb values.

Lithology of the lower Zuun-Arts Formation
In Bayan Gol, the boundary between the Shuurgat Formation and the overlying Zuun-Arts Formation is defined by a karst surface.The BAY1 section commences at the top of this karst surface and intersects a 14.1 m thick (true thickness) dolostone that preserves columnar Boxonia grumulosa stromatolites (Fig. 3).In the BAY1 section, lowermost levels in the Zuun Arts Formation are frequently obscured by alluvium, however a distinctive 90 cm interval (true thickness) of black chert (phosphatic or phosphatised chert, Bold et al., 2013, Smith et al., 2016aand Adachi et al., 2019) occurs at BAY1/42.3.Stratigraphically above the chert is a 5.2 m interval of poorly outcropping green, laminated shale.Above these levels, much of the lower Zuun Arts Formation is characterised by well-bedded, nodular, grey limestone with black siliceous balls and crusts developed on the surface.These textures do not persist in limestones intersected by the upper BAY1 section which tend to be thinly bedded and flaggy (Fig. 3).

Lithology of the upper Zuun-Arts Formation
Lowermost levels in the BAY4 section are cross-bedded grainstones with sand-sized carbonate grains cemented by microspar (Fig. 6A).Carbonate grainstones give way to oolitic limestones at BAY4-7.0 (Fig. 6B).This level also includes darker clasts of reworked oolitic limestone.Levels with fine oolites and mottled, sub-rounded intraclasts are also characterised by secondary dissolution features such as stylolites and patchy dolomitisation (Fig. 6C-D).Crossbedding, reworked carbonate grains and oolites give strong indication of a shallow-water marine depositional setting for the upper Zuun-Arts Formation.Smith et al. (2016a) defined the contact between the Zuun-Arts and Bayangol formations as a "sharp" transition between ooid grainstone and microcrystalline phosphatic shale (pictured in Smith et al., 2016a, fig. 3D).Phosphatic deposits are often associated with drowning due to major transgressive events, and "phosphatic shales" have been documented near the bases of the Zuun-Arts and Bayangol formations (Macdonald et al., 2009;Smith et al., 2016a).However, in the present study, phosphate appears to be more readily associated with the carbonates of the upper Zuun-Arts Formation (and the chert in the lower Fig. 6.Thin section photomicrographs of fabrics and textures from the upper Zuun-Arts Formation in the BAY4 section, Bayan Gol.A, BAY4/1.5, aggregation of fine carbonate grains, coated and cemented with microspar.B, BAY4/7.0,oolitic limestone (includes large, dark, oolitic intraclasts) with abundant dissolution textures.C, BAY4/10.0, mottled limestone with abundant stylolitisation and patchy dolomitisation.D, BAY4/11.5, fine oolites, rounded intraclasts and coated grains in recrystallized matrix.E, BAY4/13.0, dissolution boundary in glauconitic limestone.Glauconite grains show pale, recrystallized rinds.F, BAY4/17.5, mottled, fine micritic limestone with subcircular shadows of possible tubular fossils.
Stratigraphically lower levels in the Zuun-Arts Formation include well-developed phosphorites (Macdonald et al., 2009;Bold et al., 2016aBold et al., , 2016b;;Smith et al., 2016a), supporting the relationship between the Zuun-Arts Formation and deposition of phosphatic facies.While phosphorites are commonly interpreted as resulting from major transgressions, not all phosphatic facies may be indicative of drowned or deeper water settings.Extensive deposits of shallow water phosphorites in the Georgina Basin of Queensland and the Northern Territory of Australia are associated with carbonate textures similar to that seen in the upper Zuun-Arts Formation, in addition to the occurrence of glauconitic sediments (Southgate, 1988;Southgate and Shergold, 1991;Creveling et al., 2014).Relative proportion of phosphate in the Thorntonia Limestone in the Georgina Basin shows dramatic increase toward the upper levels immediately underlying the boundary with the Arthur Creek Formation (Southgate, 1988;Creveling et al., 2014).Similarly, phosphatic facies in the upper Zuun-Arts Formation become more frequent in the upper intervals immediately prior to deposition of the overlying Bayangol Formation, and were likely to have been deposited in a restricted shallow water palaeoenvironment conducive for development of ooid shoals, crossbedding and occasional breakup and redeposition of limestone and phosphatic material.

Lithology of the lower Bayangol Formation
The boundary between the Zuun-Arts and Bayangol formations is considered to be a major sequence boundary based on the interpretation of phosphatic shales in the lower Bayangol Formation (Lindsay et al., 1996;Smith et al., 2016a).Previous work describes the lower Bayangol Formation as a mixed carbonate and siliciclastic succession (Lindsay et al., 1996) with phosphatic shale interbedded with nodular carbonate at the base (Khomentovsky and Gibsher, 1996;Smith et al., 2016a).In contrast, we interpret phosphatic crusts, hardgrounds and stringers and reworked phosphatic pebbles as being associated with the carbonates of the underlying Zuun-Arts Formation with little to no phosphate detected in the lower Bayangol Formation.It is unclear why previous workers (e. g.Smith et al., 2016a) interpreted the "shale" in the lower Bayangol Formation as phosphatic, as no evidence has been presented to demonstrate this.Microfacies investigations and elemental mapping reveal that the lower Bayangol Formation actually represents a limestone-marl sequence and that the interbedded "shales" are of diagenetic origin (Figs 10C-E).Limestone-marl deposits like that in the lower Bayangol Formation are common in carbonate settings.They display regular alternations between carbonate and marl ("shaley") beds that appear to be the result of primary sedimentary processes oscillating between deposition of carbonate and siliciclastic muds.However, in many such sequences the marls between limestone layers are produced via carbonate dissolution during early diagenetic processes, with increased relative compaction of the marl occurring as sedimentation continues (Munnecke and Samtleben, 1996).Primary fabrics such as bedding may influence these processes and produce the rhythmic layering often seen in these deposits (Fig. 7G).Dissolution textures in the lower Bayangol Formation in BAY4/5 are obvious in thin section (Fig. 10A-C), and also via elemental mapping (Fig. 10D).Limestones in BAY5-8.0 contain only background levels of phosporous, while the supposedly phosphatic 'shales' do not contain detectable levels of phosphorous, instead showing elevation of elements like iron, silica and titanium, products likely mobilized during diagenesis (Fig. 10C, D).In addition to strong stylolitisation, microfabrics clearly show dolomitized brown marly intervals with a combination of diffuse boundaries, or boundaries that show anastomosing stylolite swarms penetrating surrounding mottled, grey micrite (Fig. 10A-E; Flügel, 2010).For the most part, these intervals lack evidence for primary siliciclastic input such as quartz grains, indicating a carbonate-dominated depositional system rather than a carbonate system with regular siliciclastic input.
In the BAY4/5 section, nodular limestones and marls are logged for 71.27 m from BAY5-2.7 to 5,10).Shelly fossils are rare in this interval, perhaps due to the pervasive development of diagenetic marls.The siliciclastic component gradually increases upsection, with occasional sandy limestone beds cropping out as prominent benches at ~50 m above the base of the section.Intervals of well-developed flat pebble conglomerates also occur from ~50-75 m above the base of the section.Planolites and Psammichnites are recovered from silty limestones, 59.86 m and 81.57m above the base of the BAY4/ 5 section (Fig 5,12), but traces tend to be rare in the lowermost Bayangol Formation.Interpretation of this part of the succession as being carbonate-dominated as opposed to mixed carbonate and siliciclastic may explain the lack of well-developed ichnofossils (including T. pedum) at these levels.Probable algal filaments are also found at some silty levels in the lower Bayangol Formation (Fig. 13).
At BAY5/111.5 (83.66 m above the base of the BAY4/5 section) there is an abrupt lithological change from limestone-marl alternations with occasional flat pebble conglomerates to massive limestone with clotted microbial textures (Figs 5, 10F,G).Microbial textures are consistent with Renalcis and Korilophyton, noted at similar levels in the Bayan Gol section by Adachi et al. (2021).Abundant dissolution-compaction features characterise these massive to thickly-bedded limestones (Fig. 10G).Grainstone textures occur with microbial clots at BAY5-131.0,103.16 m above the base of the BAY4/5 section (Fig. 10H).These give way to fine oolites and pellets coated by micritic envelopes at BAY5-149.4,121.56 m above the base of the BAY4/5 section (Fig. 10I, J).Aggregations of oolites and pellets form grapestone and carbonate lump textures at BAY5-150.1 (Fig. 10K, L).Development of pellets, oolites and grapestone textures suggests a relatively shallow water, low-moderate energy palaeoenvironment that experienced periods of nondeposition or reduced sedimentation rates (Flügel, 2010).
stromatolites (BAY5-221.2to BAY5-222.8)(Fig. 5).A thin, yellowed crust overlying these stromatolites contains abundant fine, dark, elongate structures observable in hand sample.In thin section, these structures have diffuse boundaries with granular interiors, often containing fine angular quartz (Fig. 11L, N), suggestive of burrow infills.Tubular shelly debris also occurs at BAY5-222.8, and cross sections are observable in thin section (Fig. 11M).Incorporation of siliciclastic material at BAY5-222.8 is consistent with the development of thick sandstone and shale intervals stratigraphically above this level (in a measured section complementing, but outside the scope of the present study).Fig. 10.Thin section photomicrographs of fabrics and textures from the lower Bayangol Formation in the BAY5 section, Bayan Gol.A, BAY5/1.3,yellow, mottled micritic limestone with stylolites.B, BAY5/3.3,mottled micritic limestone with rare rounded carbonate grains.Carbonate grains have diffuse boundaries.Close-set stylolites appear as dark dissolution fields.C-D, BAY5/8.0,limestone-marl interbed with yellowish, diagenetic, dolomitised pseudobed with fine grey micrite above and below.Boundaries between micrite and yellowish intervals are often diffuse, suggesting that they are secondary, rather than primary bedding.D, BAY5/8.0,elemental map of section in C shows low concentration of phosporous in the micritic intervals while diagenetic enrichment of iron, titanium and silica occurs in the marls.Previous assessment of these facies in the Bayangol Formation described "phosphatic shales" (Smith et al., 2016a).Phosphate is clearly associated (in low amounts) with the carbonates rather than the 'shales' which have no/negligible phosporous and are clearly of diagenetic origin.E, BAY5/69.5, dissolution fabrics delineating fine and granular micrite.F, BAY5/112.5,mottled micrite with clotted, microbially-mediated (Renalcis) textures.Stylolites and yellowish dissolution textures wrap around microbial fabrics.G, BAY5/117.2,fine micritic limestone with abundant stylolites and dissolution textures.H, BAY5/131.0,anastomosing stylolite swarms in intraclastic and clotted microbial limestone.I-J, BAY5/149.4,fine oolites and coated intraclasts with strong stylotisation.J, close up of I showing micritic envelopes surrounding intraclasts and collapsed oolites with dolomitized interiors.K-L, BAY5/150.1,grapestone texture; irregularly shaped intraclasts composed of reworked micritic limestone containing smaller carbonate grains.Intraclasts (occasionally recrystallized to spar) sit within a fine, light-coloured recrystallized matrix.

New interpretation of the Bayan Gol sequence
Lithological and microfacies data presented herein demonstrate that the upper Zuun-Arts Formation and the lower Bayangol Formation were both deposited in a relatively shallow, predominantly carbonate palaeoenvironment.This contrasts with previous interpretations that cite a major shift in sedimentation style at this boundary from a carbonate grainstone to mixed carbonates and siliciclastics.Sequence stratigraphic approaches differ between siliciclastic and carbonate successions (Flügel, 2010).In carbonate successions sequence boundaries can be characterised by hardgrounds, crusts and breccias that indicate transgressive flooding surfaces (Flügel, 2010).Development of phosphatic hardgrounds in the upper Zuun-Arts Formation indicates periodic sedimentary quiescence.The transition from carbonate grainstones with abundant phosphatic textures and a thick, blackened phosphatic crust at BAY4-13 (=BAY5-0) to a limestone-marl sequence at BAY5-3.3 most likely indicates a sequence boundary between the upper Zuun-Arts Formation (highstand systems tract) and the lower Bayangol Formation (transgressive systems tract).While it is difficult to determine the length of time captured within a single phosphatic hardground, continuation of carbonate-dominated sedimentation, rather than sudden influx of terrestrial detritus, in addition to the continuation of stratigraphic ranges of species of SSF across this boundary suggest that the time break at the boundary between the Zuun-Arts and Bayangol formations is likely to be minimal.

Biostratigraphy
The only skeletal fossils in the BAY1 section occur within a 2.4 m thick interval (between BAY1/104 and BAY1/107.8)and consist of poorly preserved annulated tubular fossils identified as Zuunia chimidtsereni, recently described from the basal Zuun-Arts Formation at Bayan Gol (Yang et al., 2020) and a few other poorly preserved tubular fossils (Fig. 4).The first skeletal fossils in the BAY4/5 section occur in association with phosphatic hardgrounds in the uppermost Zuun-Arts Formation (11.35 m true thickness from the base of the section, BAY4-13).Here protoconodont elements (Fig. 9A-J) occur with abundant Zuunia chimidtsereni (Fig. 9K-S).Zuunia chimidtsereni from this level are frequently preserved in three dimensions (Fig. 9K-M), sometimes with a perfectly circular cross section but are more often partly compressed with a wrinkled surface and partial detachment of subsequent growth increments (Fig. 9Q-S), suggesting soft deformation of a poorly mineralized or organic wall.As discussed by Yang et al. (2020), these tubes are closely comparable to cloudinids, common fossils from the late Ediacaran (Selly et al., 2020).Associated protoconodonts are represented by slender, recurved spine-like cones with a pointed apex.Most specimens show a well-developed median ridge on the concave side and a shallow basal furrow on the convex side (Fig. 9A-J).Material includes specimens with a relatively wide base attributable to Protohertzina anabarica and slender, laterally compressed specimens comparable with P. unguliformis.Protohertzina anabarica and P. unguliformis are usually reported as separate species (e.g., Missarzhevsky, 1973;Qian and Bengston, 1989).However, Kouchinsky et al. (2017) suggested that both forms in lower Fortunian strata are likely to belong to one apparatus and grouped them under P. anabarica (although still designating P. unguliformis-type sclerites).While both taxa have a global distribution and frequently occur together in early Cambrian SSF assemblages, Kouchinsky et al. (2017) note that the topotype material of P. unguliformis from the Tommotian-aged basal Pestrotsvet Formation is significantly younger.Herein, we follow Kouchinsky et al. (2017) and consider it most likely that protoconodont elements from the lower Bayangol Formation belong to the same apparatus and suggest that they fit within the concept of P. anabarica.However, pending formal synonymy, for biostratigraphic clarification we distinguish occurrences of P. unguliformis-type elements in our biostratigraphic range chart (Fig. 5).
In the overlying limestone-marls of the lowermost Bayangol Formation, shelly fossils are less common but specimens of Protohertzina occur at 52.58 m and 62.97 m above the base of the Bayangol Formation.The last occurrence of Zuunia is at 62.97 m above the base of the Bayangol Formation.A low diversity trace fossil assemblage is found in the clastic-influenced interval between 45-80 m and include forms comparable to Planolites and Psammichnites, common in Cambrian deposits (Figs 5,12).
In the overlying massive microbial carbonates (83.66 m to 149.87 m above the base of the Bayangol Formation), shelly fossils are initially rare but diversity increases dramatically towards the upper levels.Protohertzina anabarica (including P. unguliformis-type sclerites) ranges throughout this part of the section.The first occurrence of Anabarites trisulcatus, Cambrotubulus decurvatus and Siphogonochites triangularis are at 122.26 (BAY5/150.1)meters true thickness above the base of the BAY4/5 section, while Lopochites latazonalis, Halkieria sp. and Maikhanella multa occur at 149.16 m (BAY5/177; Figs 5, 14) and Mongolodus rostriformis and Dabashanites mirus occur at 149.86 m of the BAY4/5 section (BAY5/196.5).This fauna ranges into strata associated with the domal stromatolites at the top of the section where they occur together with the star-shaped trace fossil Spatangopsis mongolica (Fig. 15).
Fossil fauna documented here from the upper Zuun-Arts and lower Bayangol formations give both Ediacaran and Cambrian biostratigraphic signals.The lower BAY4/5 section captures a Zuunia-Protohertzina skeletal assemblage suggestive of Ediacaran-Cambrian ages, in addition to an unquestionably Cambrian-aged assemblage that includes anabaritids, siphogonochitids, Halkieria, Maikhanella and Mongolodus at stratigraphically higher levels.

Chemostratigraphy
The δ 13 C profile of the lower Zuun-Arts Formation shows gradually decreasing δ 13 C values from +3.23‰ near the base of the formation to -5.7‰ at BAY1/226.1 m, 155.8 m true thickness above the section base (Fig. 3).This overall decline in δ 13 C values is interrupted by a peak of +3.27‰ at BAY1/164.3 m, 112.1 true thickness above the section base of BAY1 (Fig. 3).
Strongly negative δ 13 C values occur in the upper Zuun-Arts Formation (-10.2‰ at BAY4/1.5, true thickness of 1.5 m above the base of the BAY4/ 5 section; Fig. 5).δ 13 C values gradually become more positive upsection, culminating in a positive peak (2.9‰ at BAY5/51.4) in the lower Bayangol Formation, 39.55 m true thickness from base (Fig. 5).δ 13 C values gradually become more negative over the next ~35 m, dropping to -5.4‰ at BAY5/98 m, 73.79 m true thickness from base (Fig. 5).δ 13 C values over the remainder of the section are negative, but exhibit a gradual climb toward 0.0‰ in the upper parts of the section (Fig. 5).

The E-C boundary at Bayan Gol
Studies over the last 25 years have generally agreed on the approximate placement on the E-C boundary in Bayan Gol.Brasier et al. (1996aBrasier et al. ( , 1996b) ) placed the E-C boundary proximal to the base of the Bayangol Formation using ichnofossil assemblages, and Smith et al. (2016a) reached a similar conclusion based primarily on chemostratigraphic data.This position of the E-C boundary has also been tentatively adopted in more recent studies (Yang et al., 2020;Adachi et al., 2021).However, in Bayan Gol, chemostratigraphic markers, the FO of key shelly fossils and the FO of the index fossil T. pedum are seemingly at odds, and recognition of the E-C boundary in Bayan Gol is reliant on proxies that are not incorporated into the current definition of the boundary at the GSSP (Voronin et al., 1982;Brasier et al., 1996b;Goldring and Jensen, 1996;Smith et al., 2016aSmith et al., , 2017;;Landing and Kruse, 2017).
Trace fossils have limited utility in the Zavkhan Basin for delineating the base of the Cambrian due to the delayed appearance of T. pedum (the FO of the species is in 'unit 20' in Bayan Gol, ~250 m above the base of the Bayangol Formation; Goldring and Jensen, 1996), highlighting the importance of SSF biostratigraphical data.Collection of new SSF data from these sections is crucial as previous work is either fraught with mapping and correlation errors (described above, Brasier et al., 1996b), or does not document fossil taxa and their occurrences in adequate detail (Smith et al., 2016a).Smith et al. (2016a) collated a vast amount of palaeontological data from Mongolian sections measured by previous workers (e.g., Korobov and Missarzhevsky, 1977;Voronin et al., 1982;Endonzhamts and Lkhasuren, 1988;Zhegallo and Zhuravlev, 1991;Dorjnamjaa et al., 1993;Brasier et al., 1996b;Esakova and Zhegallo, 1996).However, uncertainties surrounding the provenance of many of these samples, and developments in shelly fossil palaeobiology that have prompted taxonomic revisions for key taxa render old biostratigraphic data unreliable.
The marriage of shelly fossil biostratigraphic data and δ 13 C chemostratigraphic data is key for putting regional Cambrian successions in a global context.The first skeletal faunas (including the protoconodont Protohertzina anabarica) appear 9.85 m above the nadir of the negative excursion in the Zuun-Arts Formation (2.12 m below base of the Bayangol Formation).The occurrence of Protohertzina in combination with the cloudinid Zuunia chimidtsereni (an Ediacaran-Fortunian taxon; Yang et al., 2020) in the upper Zuun-Arts Formation corroborates earlier suggestions that the strongly negative δ 13 C values in the upper Zuun-Arts Formation represent the BACE (-10.2‰ at 10.46 m below the base of the Bayangol Formation, Fig. 5).The BAY4/5 section was measured through continuous strata on a single fault block and therefore only captures the nadir and upper part of the excursion, while the BAY1 section likely captures the lower part of the BACE (Fig. 3).
The BACE is recognized globally and has been used in the absence of a reliable FO of T. pedum as a proxy for defining the base of the Cambrian (see Zhu et al., 2019 and more discussion below).The close temporal relationship between the FO of Protohertzina and the nadir of the BACE in the upper Zuun-Arts Formation indicates that the E-C boundary in Bayan Gol occurs in the upper Zuun-Arts Formation.Herein, we place the E-C boundary at 10.46 m below the contact with the overlying Bayangol Formation, corresponding with the nadir of the BACE (Fig. 5).

Chronostratigraphic markers for the E-C boundary
Identification of the E-C boundary in Bayan Gol herein (i.e.coinciding with the nadir of the BACE) has been achieved using proxies not present at the Cambrian GSSP at Fortune Head in Newfoundland, preventing direct correlation between Bayan Gol and the global stratotype section.Similar situations are encountered on most palaeoncontinents and clearly demonstrate the profound difficulties with reliably defining the base of the Cambrian around the world (Steiner et al., 2007(Steiner et al., , 2020;;Yang et al., 2014aYang et al., , 2016;;Zhu et al., 2017Zhu et al., , 2019;;Kouchinsky et al., 2017;Yang and Steiner, 2021).We address the advantages and shortcomings of the Cambrian GSSP at Fortune Head and explore additional chronostratigraphic markers that may provide a practical solution to problems that plague global correlation of one of the most significant chronostratigraphic boundaries in the geological timescale.

The current GSSP definition
The E-C boundary is defined by the first appearance of the ichnofossil Treptichnus pedum (Narbonne et al., 1987;Brasier et al., 1994;Landing, 1994).At the stratotype section at Fortune Head in Newfoundland the Cambrian GSSP is marked at a point 2.3 m above the base of Member 2A (now incorporated in the Quaco Road Member) of the Chapel Island Formation (Fig. 16; Narbonne et al., 1987;Brasier et al., 1994;Landing, 1994;Landing, 1996).This horizon additionally marks the base of the T. pedum Ichnofosssil Assemblage Zone (Narbonne et al., 1987;Babcock et al., 2014).Global correlation of the GSSP can be achieved by either identifying the first appearance of T. pedum or other biostratigraphic ichnofossil indicators of the T. pedum Ichnofossil Assemblage Zone, such as Skolithos annulatus, Arenicolites sp., Monomorphichnus spp., Conichus conicus, Phycodes sp. or Helminthopsis tenius that all occur within 2 m of the GSSP (see Landing, 1994;Peng et al., 2012;Babcock et al., 2014).
On the other hand, the deficiencies of employing the FAD of T. pedum to define the base of the Cambrian have been noted repeatedly (Peng and Babcock, 2011;Babcock et al., 2014;Zhu et al., 2019).Critics have questioned its stratigraphic distribution, uncertain taxonomic history and behavioural significance (Dzik, 2005;Vannier et al., 2010;Peng et al., 2012;Babcock et al., 2014;Buatois, 2018).The efficacy of T. pedum as a chronostratigraphic marker was further undermined when it was discovered 4.41 m below the E-C boundary in the section at Fortune Head where the GSSP was established (Gehling et al., 2001).Some saw this as a major problem (e.g., Peng and Babcock, 2011), while others considered this occurrence to fall within acceptable confidence intervals (Landing et al., 2013).To compensate for the discrepancy, Landing et al. (2013) suggested redefining the Cambrian GSSP to correspond with the base of the T. pedum Ichnofossil Assemblage Zone rather than the FAD of T. pedum itself.The base of the T. pedum Ichnofossil Assemblage Zone is defined immediately above the last occurrences of the enigmatic Harlaniella podolica Sokolov, 1972 andPalaeopascichnus delicatus Palij, 1976 with the first occurrence of T. pedum now slightly below this boundary (Landing et al., 2013).This revised interpretation has been met with some resistance by critics that suggest this only further complicates the identification of the base of the Cambrian (Babcock et al., 2014;Zhu et al., 2019).
Perhaps the most significant complication when employing T. pedum for global correlation is that of facies dependence (Geyer and Uchman, 1995;Gehling et al., 2001;Geyer, 2005;Babcock et al., 2014;Zhu et al., 2019).Many trace fossils (including T. pedum) are predominantly confined to siliciclastic rocks, and although it has been noted that T. pedum occurs in a range of lithofacies (Landing et al., 2013;Buatois et al., 2013;Buatois, 2018) these do not typically include carbonate settings.This lithofacies dependence dramatically reduces the application of T. pedum as a proxy for the base of the Cambrian, as in many key regions this interval captures predominantly carbonate facies, e.g., China, Kazakhstan, Siberia and Mongolia.Facies restriction may explain the delayed appearance of T. pedum in some E-C sections (Buatois et al., 2013;Buatois, 2018) such as Bayan Gol herein.Likewise, traces associated with the T. pedum Ichnofossil Assemblage Zone such as Skolithos annulatus or Arenicolites sp. are also facies dependent.Other ichnotaxa such as H. podolica and P. delicatus have limited palaeogeographical distribution (Zhu et al., 2019) and also appear to be largely restricted to siliciclastic settings and are hence unlikely to be reliable markers for the base of the T. pedum Ichnofossil Assemblage Zone, or robust alternative markers for the base of the Cambrian.

Where is T. pedum a useful marker?
Nevertheless, the use of T. pedum and the Fortune Head GSSP has been staunchly defended (Gehling et al., 2001;Landing et al., 2013;Geyer and Landing, 2017;Buatois, 2018).Recognition of the E-C boundary in Namibia, northwestern Canada, the Great Basin, Norway, central England and Australia is often presented as evidence that T. pedum is an effective global marker (Gehling et al., 2001, p. 214;Landing et al., 2013, p.145).However, it is clear that the occurrences of T. pedum in some of these locations do not closely correlate with the Cambrian GSSP at Fortune Head.In the Arrowie Basin in South Australia, the first occurrence of T. pedum occurs approximately 200 m above the proposed E-C boundary (Jensen et al., 1998;Zang et al., 2007;Jago et al., 2006Jago et al., , 2020) ) and 50 m above the proposed boundary in the Amadeus Basin in central Australia (McIlroy et al., 1997).The boundary in both basins is disputed, complicated by the presence of a major unconformity (Zang et al., 2007;Jago et al., 2006Jago et al., , 2020)).Nevertheless, as the first occurrence of T. pedum is above Treptichnus coronatum in the Arrowie Basin (Jensen et al., 1998;Zang et al., 2007) and above Rusophycus in the Amadeus Basin (McIlroy et al., 1997), the FO of T. pedum in Australia clearly does not define the E-C boundary (McIlroy et al., 1997;Gehling, 2000;Jago et al., 2006Jago et al., , 2012;;Gehling and Droser, 2012;Shahkarami et al., 2020).
In Namibia, traces that have similar morphology and complexity to T. pedum such as Streptichnus narbonnei occur in the upper Spitskop Member, 14 m below the unconformable contact with the overlying Nomtsas Formation (Fig. 16; Linnemann et al., 2019, fig. 3).However, the first unambiguous occurrence of T. pedum is in the lower Nomtsas Formation, where the E-C boundary is currently placed (Fig. 16; Germs, 1972;Wilson et al., 2012;Linnemann et al., 2019;Bowyer et al., 2022).An ash bed in the Nomtsas Formation provides a maximum age of 538.8 Ma for T. pedum (Bowyer et al., 2022), however, this has been challenged by recent suggestions that the boundary may be >1 m.y.younger than currently recognized (Nelson et al., 2022).In central England, T. pedum has not been reported and correlation with the GSSP is based on the presence of the ichnotaxon Teichichnus (Bland and Goldring, 1995;McIlroy et al., 1998;Landing et al., 2013).McIlroy et al. (1998) correlated the Teichichnus bearing beds in the Swithland Formation with the Teichichnus ichnozone of Newfoundland (Landing, 1992), that occurs some 250 m above the GSSP (McIlroy et al., 1998).
The application of T. pedum to reliably recognize the E-C boundary works best in predominantly siliciclastic successions (Fig. 16), such as Finnmark in northern Norway (Högström et al., 2013), southeastern Poland (Moczydłowska, 2008) and some regions in North America (e.g., Narbonne et al., 1994;Corsetti and Kaufman, 1994;Corsetti andHagadorn, 2000, 2003;Smith et al., 2016b;Tarhan et al., 2020).Trace fossils provide the main age constraint on the succession in Finnmark (Fig. 16).Here, Palaeopascichnus occurs below T. pedum, and the first occurrence of T. pedum is associated with trilobed traces indicative of the T. pedum Ichnofossil Assemblage Zone (Högström et al., 2013;Jensen et al., 2018).These data suggest that the E-C boundary is most likely in the upper portion of the Manndrapererlva Member of the Stáhpoógieddi Formation (Fig. 16; Högström et al., 2013;Jensen et al., 2018).However, correlation on a broader scale is challenging as complementary data is not available (e.g., chemostratigraphic or skeletal fossil markers).On the Lublin slope in southeastern Poland, the FO of T. pedum is in the lower Mazowsze Formation (Fig. 16; Strauss et al, 1997;Moczydłowska, 2008).However, the E-C boundary on the Lublin slope is not defined by the ichnotaxon but rather the FO of Cambrian acritarch species in the uppermost part of the underlying Włodawa Formation (Fig. 16; Strauss et al., 1997;Moczydłowska, 2008).Although acritarchs have been employed on the Lublin slope to delineate the E-C boundary, the temporal and spatial relationship of acritarch assemblages with the FO of T. pedum on a global scale has not been evaluated.Consequently the E-C boundary on the Lublin slope remains under discussion.
In the Great Basin (Corsetti and Kaufman, 1994;Corsetti and Hagadorn, 2000, 2003, Ahn et al., 2012;Loyd et al., 2012;Smith et al., 2016b;Tarhan et al., 2020), Mexico (Hodgin et al., 2021) and the Mackenzie Mountains of northwest Canada (Narbonne et al., 1994;Narbonne and Aitken, 1995;Carbone and Narbonne, 2014) there is a close association between the first occurrence of T. pedum with the upper part (post nadir) of the BACE (Fig. 17).However, it is not possible to establish whether these sections may be correlated with the stratotype due to the lack of chemostratigraphic or skeletal biostratigraphic data from the siliciclastic-dominated Fortune Head.Although negative δ 13 C org values have been recorded from the Chapel Island Formation (Fig. 16; Hantsoo et al., 2018), these values represent organic carbon and not inorganic carbonate carbon isotopes.The relationship between carbonate and organic δ 13 C values is not clear (e.g., Kump and Arthur, 1999;Oehlert and Swart, 2014).Hence, chemostratigraphic excursions  T. Topper et al. in organic carbon (δ 13 C org ) isotope curves obtained through the Newfoundland GSSP section (Fig. 16; Hantsoo et al., 2018) may not be equivalent to the BACE observed in carbonate carbon (δ 13 C carb) isotope curves elsewhere. .Using the base of the T. pedum Ichnofossil Assemblage Zone rather than T. pedum, for correlation, as proposed by Landing et al. (2013) does not enhance correlation, as neither Harlaniella or Palaeopascichnus have been documented from these sections (Narbonne and Aitken, 1995;Corsetti and Hagadorn, 2000, 2003, Carbone and Narbonne, 2014;Smith et al., 2016b;Kolesnikov, 2019;Tarhan et al., 2020;Hodgin et al., 2021).

T. pedum is an inadequate correlative marker
The combination of facies dependence and insufficient correlative markers at the stratotype section places serious limitations on the ability to correlate globally using either the FAD of T. pedum or the base of the T. pedum Ichnofossil assemblage Zone from the GSSP on Fortune Head.It is clear that while T. pedum may have proven its utility in identifying Lower Cambrian strata around the globe, as an unambiguous global marker for the E-C boundary it is inadequate.
Stratigraphic overlap of what was previously considered exclusively Ediacaran and Cambrian taxa complicates the application of SSF to define the base of the Cambrian, and the discovery of Anabarites in terminal Ediacaran strata appears to have given rise to skepticism regarding the biostratigraphic utility of skeletal taxa in defining the E-C boundary (see Landing et al., 2013;Zhu et al., 2017).However, the extension of fossils into unfamiliar stratigraphic territory is in reality restricted to only a handful of taxa, notably cloudinids and other tubular forms like Cambrotubulus and Anabarites.Even A. trisulcatus, which is now frequently considered an Ediacaran taxon has only been convincingly shown in Ediacaran strata (based on our boundary definition) at the top of the Dengying Formation in Lijiagou by Cai et al. (2019) where it occurs just below the nadir of the BACE.Siberian records of A. trisulcatus in Ediacaran strata are based on misinterpreted δ 13 C profiles (Zhu et al., 2017;Bowyer et al., 2022; see discussion below), inaccurate definition of the lower boundary of the A. trisulcatus Zone (Rogov et al., 2015;Nagovitsin et al., 2015;Kouchinsky et al., 2017) and U-Pb zircon dates (Bowring et al., 1993) that likely need to be revisited, given the advances in zircon preparation and recalibration of standards (Schmitz, 2012;Kouchinsky et al., 2017).The unclear stratigraphic position of anabaritids and cloudiniids indicates that they should be avoided as primary biostratigraphic tools for the base of the Cambrian.Furthermore, these tubular forms are secondarily phosphatized (Yang et al., 2020) and are consequently highly facies dependent and are therefore unreliable as index fossils (Pruss et al., 2018;Freeman et al., 2019;Jacquet et al., 2019).

Where is Protohertzina a useful marker?
Protohertzina is used to indicate basal Cambrian strata globally (Steiner et al., 2007;Geyer, 2019).For instance, the first occurrence of Protohertzina has been employed to help place the E-C boundary in the Alborz Mountains of Iran (Fig. 18; Hamdi et al., 1989;Shahkarami et al., 2017;Fig. 18;Shahkarami et al., 2020) and the Mackenzie Mountains in northwest Canada (Fig. 17; Narbonne et al., 1994;Pyle et al., 2006).Protohertzina forms a key element in the oldest shelly fossil assemblages in Siberia and South China-regions that host perhaps the most extensive and richly fossiliferous Fortunian strata in the world (Figs 19,20).
In Siberia, Protohertzina anabarica is part of the Anabarites trisulcatus Zone, typically first occurring above the FO of A. trisulcatus.It has been documented across the Siberian Platform, including from the Syhargalakh Formation of the Olenek Uplift, the Ust'Yudoma Formation in eastern Siberia and the Nemakit-Daldyn Formation in the Anabar Uplift (Figure 19; see Kouchinsky et al., 2012Kouchinsky et al., , 2017)).In South China, the first occurrence of P. anabarica (or the likely junior synonym P. unguliformis) defines the base of the Anabarites trisulcatus -Protohertzina anabarica Assemblage Zone (Steiner et al., 2007;Yang et al., 2014a).The Anabarites trisulcatus -Protohertzina anabarica Assemblage Zone has been identified from the majority of carbonate platforms in South China, such as the Maidiping Formation in central Sichuan, the lower Kuanchuanpu Formation in south Shaanxi, the Yanjiahe Formation in Hubei and the Zhongyicun Member of the Zhuijaqing Formation across the Yangtze Platform (Steiner et al., 2007(Steiner et al., , 2020;;Guo et al., 2014;Yang et al., 2014aYang et al., , 2016;;Liu et al., 2020;Yang and Steiner, 2021;Fig. 20).The base of both the Anabarites trisulcatus Zone in Siberia and the Anabarites trisulcatus-Protohertzina anabarica Zone in South China are considered to approximately coincide with the base of the Cambrian in their respective regions (Steiner et al., 2007(Steiner et al., , 2020;;Zhu et al., 2017;Kouchinsky et al., 2017;Geyer, 2019).

Protohertzina and the BACE
One advantage of employing skeletal fossils as tools for global correlation is that as they predominantly occur in carbonate rocks, chemostratigraphic records can also be obtained, providing an auxiliary means to strengthen correlation.The integration of biostratigraphic and chemostratigraphic data has become increasingly applied in Cambrian studies (Betts et al., 2018;Steiner et al., 2020) and many key stratigraphic sections that were initially sampled for skeletal taxa have now been studied for chemostratigraphic purposes (Li et al., 2009(Li et al., , 2013;;Smith et al., 2016a;Betts et al., 2018).Consequently, the relationship between prominent perturbations in the δ 13 C profile and the first appearances and stratigraphic ranges of key skeletal taxa are becoming clear.
In many successions around the globe the first occurrence of Protohertzina is closely associated with the BACE.In eastern Siberia, the first occurrence of Protohertzina is approximately 60-75 m above the nadir of the BACE (excursion 'N' of Braiser et al., 1993) in the Ust'-Yudoma Formation, approximately 82 m below the contact with the Pestrotsvet Formation (Khomentovsky and Karlova, 1992;Braiser et al., 1993;Kouchinsky et al., 2007Kouchinsky et al., , 2012)).In the western Anabar region (Kotujkan section) in northern Siberia Protohertzina has been recovered from the basal Manykaj Formation (Khomentovsky andKarlova, 1992, 1993), approximately 40 m above the nadir of the BACE (Kaufman et al., 1996;Kouchinsky et al., 2017).In the sedimentary successions of the Olenek Uplift in northeastern Siberia, Protohertzina first occurs in the upper Syhargalakh Formation (lower Kessyusa Group), 15-20 m above a negative shift in δ 13 C values (Knoll et al., 1995;Pelechaty et al., 1996) that may represent the BACE (see Kouchinsky et al., 2017 for discussion).The first occurrence of Protohertzina in this region coincides with the first occurrence of T. pedum and the base of the Cambrian has been placed at this level as a result (Rogov et al., 2015;Nagovitsin et al., 2015;Kouchinsky et al., 2017).
In South China, at the Xiaotan and Laolin sections of Yunnan, the FO of Protohertzina has been reported approximately 80 m and 40 m respectively above the nadir of the BACE (Fig. 20; Luo et al., 1991;Li and Xiao, 2004;Li et al., 2009Li et al., , 2013;;Yang et al., 2014a).In Hubei, Protohertzina has been documented from unit 2 of the Yanjiahe Formation (Chen, 1984;Guo et al., 2014) approximately 5-6 m above the nadir of the BACE (Steiner et al., 2020).In the Mackenzie Mountains of northwestern Canada, Protohertzina first occurs towards the top of the Ingta Formation (Conway and Fritz, 1980;Narbonne and Aitken, 1995) approximately 150m above the nadir of the BACE that occurs in the basal-mid beds of the Ingta Formation (Narbonne et al., 1994).In southwestern Mongolia the first occurrence of Protohertzina is 10.46 m above the nadir of the BACE in the uppermost Zuun-Arts Formation, documented herein (Figs. 5,18).
Protohertzina has also been documented in close association with the BACE in the Kazakh terrane and Iran, however the relationship is less clear in these regions, as biostratigraphic and chemostratigraphic data were collected at different times and from different localities (Missarzhevsky and Mambetov, 1981;Yang et al., 2016;Stammeier et al., 2019).The lack of illustrated fossil specimens and questionable taxonomy has also hindered precise correlation of these sections.In the Kazakh terrane the first occurrence of Protohertzina may be in the uppermost Kyrshabakty Formation (initially reported by Missarzhevsky and Mambetov, 1981, but not confirmed by later workers) or the lowermost Chulaktau Formation (Yang et al., 2016).These occurrences roughly coincide with the nadir of the BACE that is interpreted to occur close to the boundary between these two formations (Stammeier et al., 2019).
In Iran, the first undisputed specimens of P. anabarica occur in the Middle Dolomite Member of the Soltanieh Formation, just above the first occurrence of T. pedum in the uppermost Lower Shale Member (Hamdi et al., 1989;Shahkarami et al., 2017;Devaere et al., 2021).Hamdi et al. (1989) however, reported questionable fragments of Protohertzina sp. in the uppermost Lower Dolomite Member of the Soltanieh Formation together with Hyolithellus sp., Rugatotheca sp., and morphs of Olivooides.None of these taxa were illustrated and their occurrence in the Lower Dolomite Member, although accepted, appears not to have been confirmed by later workers (Shahkarami et al., 2017;Devaere et al., 2021).The nadir of the BACE in the Valiabad Section in North Iran is interpreted to occur close to the contact of the Lower Dolomite and Shale members (Kimura et al., 1997).However, pinpointing the exact nadir is difficult due to inadequate sampling (presumably due to the dominance of shale in particular intervals) and the presence of severe fluctuations of δ 13 C values (-15‰) in the Lower Shale Member (Fig. 18).
Association with the BACE across palaeocontinents supports a near-synchronous distribution of Protohertzina on a global scale.Importantly, in successions where reliable δ 13 C data is available, the FO of Protohertzina is always above the nadir of the BACE (closest association of ~5 m in China [Steiner et al., 2020]), with no convincing reports from older strata.The documentation of Protohertzina specimens from the Ust'-Yudoma Formation in the Kyra-Ytyga River, eastern Siberia (Zhu et al., 2017) from horizons that were interpreted as Ediacaran are likely to be erroneous.The Ediacaran age of these strata was based predominantly on a δ 13 C profile that displayed a slight decline of δ 13 C values (from 1.43% to -0.65%) towards the top of the Ust'-Yudoma Formation.Zhu et al. (2017) interpreted this chemostratigraphic curve as representing the onset of the BACE and hence the base of the Cambrian (Zhu et al., 2017(Zhu et al., , 2019;;Zhuravlev and Wood, 2018;Wood et al., 2019).Similarly, Bowyer et al. (2022) suggest that the Anabarites trisulcatus -Protohertzina anabarica Zone and the lower boundary of the Nemakit-Daldynian is Ediacaran in age (pre-BACE).This conclusion was based on the presence of Cambrotubulus decurvatus (not Anabarites trisulcatus) in the Turkut Formation (Rogov et al., 2015) and the 542.8+1.3Ma radiometric age (Bowring et al., 1993) of a volcanic breccia within the unconformably overlying Syhargalakh Formation (Rogov et al., 2015).A consequence of the interpretation of Bowyer et al. (2022) is that the first global appearances of not only anabaritids, but also a menagerie of Fig. 16.Stratigraphy and chemostratigraphy of Ediacaran-Cambrian successions from Avalonia, Baltica and Gondwana.Occurrences of key ichnotaxa and skeletal fossils are indicated.Newfoundland section compiled with data obtained from Narbonne et al., 1987;Landing, 1994;Gehling et al., 2001;Hantsoo et al., 2018.Finnmark section modified from Högström et al., 2013.Poland section compiled with data obtained from Compston et al., 1995;Strauss et al., 1997;Moczydłowska, 2008.Namibia section modified from Linnemann et al., 2019;Bowyer et al., 2022.It is important to note that the isotope trends for both Newfoundland and Poland are sedimentary organic carbon δ 13 C isotope records, not carbonate δ 13 C isotope records.Symbols shown are also utilized for succeeding figures.
Correlations of Bowyer et al. (2022) remain unsubstantiated, as the nadir of the BACE is not preserved in the Kyra-Ytyga section and δ 13 C profiles through the Ust'-Yudoma Formation at other Siberian localities display a markedly different signal (Braiser et al., 1993).For example, a δ 13 C profile through the lower-middle Ust'-Yudoma Formation at the Aldan River (Magaritz et al., 1986;Braiser et al., 1993Braiser et al., , 1994;;Kouchinsky et al., 2007;Varlamov et al., 2008) shows a rising trend through the entire Ust'-Yudoma Formation from negative values of ca.-4.5‰ (excursion 'N' in Kouchinsky et al., 2017), reaching peak values of ca.3.5‰ (excursion 'I') in the upper Ust'-Yudoma Formation (Fig. 19).The negative 'N' excursion in the lower Ust'-Yudoma Formation (Magaritz et al., 1986;Braiser et al., 1993Braiser et al., , 1994;;Kouchinsky et al., 2007;Varlamov et al., 2008) has previously been correlated with the BACE (Fig. 19) suggesting that the E-C boundary occurs at a much lower stratigraphic level in the Ust'-Yudoma Formation than presented by Zhu et al. (2017).In contrast, the BACE likely occurs approximately 185 m below the base of the Pestrotsvet Formation in the Dvortsy section (Braiser et al., 1993;Kouchinsky et al., 2017).Although Siberia hosts one of the most important E-C successions in the world, the majority of the stratigraphic sections discussed herein have not been recently investigated and likely require detailed reassement (e.g.data presented in Zhu et al. [2017] was based on unpublished material collected in 1981).Additionally, it is possible that the radiometric date from the volcanic breccia within the lower Syhargalakh Formation cited by Bowyer et al. (2022) as evidence for an Ediacaran age may be refined by future high precision CA-ID-TIMS analyses (Kouchinsky et al., 2007;Bowyer et al., 2022).

Problems with Protohertzina
Although Cambrian body fossils of chaetognaths (the organism that Protohertzina most likely represents) have been recovered from shale and mudstones (e.g., Burgess Shale and Chengjiang Lagerstätten; Shu et al., 2017;Caron and Cheung, 2019), cherts (Braiser and Singh, 1987;Yang et al., 2014b) and specimens questionably assigned to Protohertzina have been found in siliciclastic deposits (McIlroy and Szaniawski, 2000), the taxon is most frequently recovered from carbonate deposits via acid dissolution.This dependence on carbonate facies places limitations on its utility as a tool for global correlation.Restricted environmental tolerance explains its delayed appearance in E-C successions that are predominantly siliciclastic.For example, in Newfoundland, the FO of P. anabarica is within the upper Member 4 of the Chapel Island Formation (Landing et al., 1989), hundreds of metres above the FO of T. pedum.In South Australia, Protohertzina does not appear until the Micrina etheridgei Zone (Terreneuvian Stage 2, Betts et al., 2017bBetts et al., , 2018)).Nevertheless, the prevalence of Protohertzina in carbonate facies provides the opportunity to couple its biostratigraphic signal with δ 13 C chemostratigraphy and employ a multi-proxy approach to dating and correlation not possible in siliciclastic strata.

The advantages of δ 13 C chemostratigraphy
As doubts over the utility of trace and skeletal fossils endure, carbon isotope chemostratigraphy has rapidly become the tool of choice for high-resolution correlation in the Cambrian (Braiser et al., 1993(Braiser et al., , 1994(Braiser et al., , 1996b;;Maloof et al., 2010aMaloof et al., , 2010bMaloof et al., , 2005;;Ishikawa et al., 2008Ishikawa et al., , 2014;;Smith et al., 2016a;Chang et al., 2017;Ren et al., 2017;Schmid, 2017;Betts et al., 2018;Zhu et al., 2019).Carbonate rock successions record variations in the 13 C/ 12 C ratio of ancient ocean waters over time.The world's oceans are mixed on relatively short timescales, hence inorganic δ 13 C peaks and nadirs can provide a method for correlating contemporaneous carbonate strata globally.Carbon isotopes are also relatively resistant to major diagenetic processes (Saltzman et al., 2012) and in sufficiently continuous rock successions the resulting isotopic signature tends to remain recognizable despite any potential local diagenetic overprinting (Weissert et al., 2008;Kouchinsky et al., 2017).

Criticisms of δ 13 C chemostratigraphy
Despite the increasing reliance on chemostratigraphy for correlation of lower Cambrian rocks worldwide, the reliability and accuracy of carbon isotope chemostratigraphy has been questioned.Main criticisms of the technique are 1) that carbon isotopes can be altered by diagenetic processes (Higgins et al., 2018;Ahm et al., 2018;Bold et al., 2020;Nelson et al., 2021) and 2) that δ 13 C values in carbonate rocks reflect local environmental conditions rather than global values (Holmden et al., 1998;Swart and Eberli, 2005;Swart, 2008;Swart and Kennedy, 2012;Geyman andMaloof, 2019, 2021).Recent work on modern carbonate platforms has shown that δ 13 C values from carbonate marine sediments can vary significantly across shelf-to-slope transects, indicating that local platform processes can decouple the δ 13 C of shallow carbonates from global ocean chemistry (Swart et al., 2009;Geyman andMaloof, 2019, 2021).Absolute δ 13 C values can also be affected during burial and subsequent interaction with circulating fluids and it has been shown that dolomitzation can have a profound effect on the carbon isotopic records of carbonates (Higgins et al., 2018;Nelson et al., 2021;Bold et al., 2020).Exploring the extent to which local platform conditions and diagenesis alter the δ 13 C of bulk limestone from the primary global seawater signal is the focus of much research, and frameworks to interpret carbon isotope excursions in the geological record are being continually developed (Holmden et al., 1998;Swart and Eberli, 2005;Swart, 2008;Swart and Kennedy, 2012;Higgins et al., 2018;Geyman andMaloof, 2019, 2021;Bold et al., 2020).Oxygen isotope chemostratigraphy (δ 18 O) has been frequently used in concert with carbon isotopes to detect alteration (Marshall, 1992;Melim et al., 2001;Higgins et al., 2018).More recently, Ca and Mg isotopes in addition to traceelement ratios (such as Sr/Ca and Mg/Ca) have been employed to identify both primary mineralogy and diagenetic alteration (Higgins et al., 2018;Jones et al., 2020;Bold et al., 2020).Despite these advancements, the issue of overprinting and the role that local conditions and diagenesis plays in shifting the primary global seawater signal is often difficult to fully constrain.
These concerns have been highlighted in Cambrian strata with anomalous variation in chemostratigraphic trends that may have been altered by local aquafacies and/or experienced reworking or reoxidation of light organic carbon sources (Kouchinsky et al., 2017;Steiner et al., 2020;Yang and Steiner, 2021).For example, the Ust'-Yudoma Formation on the Siberian Platform exhibits a rising trend from negative values (-4.5‰; considered the BACE) to a positive excursion of 3.5‰ in its T. Topper et al. upper part (Fig. 19) (Magaritz et al., 1986;Braiser et al., 1993Braiser et al., , 1994;;Kouchinsky et al., 2007;Varlamov et al., 2008).However, in the nearby Kyra-Ytyga River section (Zhu et al., 2017), the BACE is not preserved and the Ust'-Yudoma Formation instead shows a plateau of positive δ 13 C values that led Zhu et al. (2017) to interpret the strata as Ediacaran.In Namibia, δ 13 C values of carbonates plateau at ~1‰ throughout the Spitskopf Member (Fig. 16; Linnemann et al., 2019), despite dates from ash beds overlapping with dates of tuffaceous material in the La Ciénga Formation in Mexico where δ 13 C values at the nadir of the excursion reach -7.5‰ (Fig. 17; Hodgin et al., 2021).Hodgin et al. (2021) suggested that the potential presence of local, isotopically distinct pools of dissolved inorganic carbon could explain the absence of the BACE in Namibia, however Bowyer et al. (2022) has argued that the positive δ 13 C profile of the Swartpunt section does not directly correlate with the BACE.
Beyond grappling with potential alteration in the primary δ 13 C signal, correlation of chemostratigraphic profiles relies on being able to accurately align positive and negative excursions.Peaks and nadirs may be indistinct, and identification may be impossible in sequences where excursions are omitted due to tectonic truncations, hiatuses or the reworking of strata (Kouchinsky et al., 2017;Steiner et al., 2020).Such difficulties may be minimized by combining carbon isotope stratigraphy with biostratigraphic tools that would place a carbon isotope excursion within a specific biozone or in association with the FAD of a particular taxon.Although this has been strongly advocated (see Betts et al., 2018;Steiner et al., 2020) it is still not standard practice and chemostratigraphy continues to be utilized as a stand-alone proxy for regional and global correlation (Maloof et al., 2010a(Maloof et al., , 2010b;;Smith et al., 2016a).In the absence of robust supporting data, miscorrelations between carbon isotope trends has created controversial age assessments (e.g., Brasier et al., 1996b;Smith et al., 2016aSmith et al., , 2017;;Landing and Kruse, 2017;Bowyer et al., 2022) and resulted in temporal scattering of the FADs of key Cambrian skeletal taxa (see Steiner et al., 2020 for discussion) There have been recent attempts to devise quantitative methods to identify Cambrian isotope peaks (Hay et al., 2019;Sinnesael et al., 2021).Such models tend to require a priori assumptions regarding stratigraphic positions of particular curves in (what are assumed to be) complete successions, and must reliably account for numerous variables such as sedimentation rates and diagenetic effects on isotopes and fossil taxa.Hence, the applicability of such methods is yet to be demonstrated, and the standard approach relies on qualitative assessments.Consequently, the reliability and accuracy of chemostratigraphic correlation relies heavily on coupling with other robust proxies, such as biostratigraphy and/or absolute radiometric dates.

Correlating a chemostratigraphic curve
Despite these challenges, it is clear that global δ 13 C trends are widely and repeatedly detected, and in carbonate rocks straddling the E-C boundary values indicating the BACE regularly occur.However, while chemostratigraphic excursions are widely used for correlation, the point on the curve at which to place the boundary is often unclear.Some use the peak or nadir (Peng et al., 2012;Narbonne et al., 2012;Cai et al., 2019), and others use the 'turning point' of the excursion where values begin to increase or decrease (Zhu et al., 2017(Zhu et al., , 2019)).Turning points in chemostratigraphic curves are difficult to recognise, and so the peak or the nadir of an excursion is the most logical position to place a temporal marker.The misinterpretation of the δ 13 C profile by Zhu et al. (2017) in the Ust'-Yudoma Formation highlights the difficulties of utilizing the onset of the excursion as a marker.In this case, there is uncertainty surrounding whether decreasing δ 13 C values indicate commencement of the BACE or simply represent minor isotopic fluctuations, or a change in signal influenced by local or regional factors.
Currently there is no standard method for defining an isotope excursion (Smith et al., 2015).However, utilizing the most distinctive feature of the δ 13 C excursion-namely the peak or nadir-is likely to be the most straightforward way to define and correlate it.In the BACE, and other negative excursions the nadir would be defined as the horizon that yielded the lowest δ 13 C values in an excursion profile.If consecutive samples yielded the exact same δ 13 C value then the oldest sample in the succession would represent the nadir of the excursion.

Recognizing the E-C boundary
Less than ten years ago, Landing et al., 2013, p. 145) declared that "the existing basal Cambrian GSSP can be readily sustained" and that "there were no strong reasons to revisit what had been an almost 30 year-long debate".However, despite this resolute attitude global correlation of the E-C boundary continues to be frustrated by the inadequacies of T. pedum (or the base of the T. pedum Ichnofossil Assemblage Zone) as a proxy for the base of the Cambrian.
A GSSP is intended to be a stable, precise horizon for global correlation.Despite this, the base of the Cambrian in numerous sections cannot be correlated using T. pedum, even if it is present.In the BAY4/5 section in Bayan Gol for example, T. pedum occurs 250 m above the base of the Bayan Gol Formation (Goldring and Jensen, 1996).However, the E-C boundary has never been positioned here, as this is clearly a considerable distance above the most likely stratigraphic location of the E-C boundary (Brasier et al., 1996a(Brasier et al., , 1996b;;Goldring and Jensen, 1996;Smith et al., 2016a;Yang et al., 2020;Steiner et al., 2020;herein).Biostratigraphic (FO of P. anabarica) and chemostratigraphic evidence indicates that the base of the Cambrian (corresponding with the nadir of the BACE) should be placed in the upper Zuun-Arts Formation, just below (10.46 m below, in our study) the contact with the overlying Bayangol Formation (Smith et al., 2016a;Adachi et al., 2019;Yang et al., 2020;Steiner et al., 2020;herein).However, this boundary must be considered informal, as neither the first occurrence of Protohertzina nor the presence of the BACE provides any direct association with the GSSP on Fortune Head.
Despite increased discussion over the last few years concerning how to best approach this dilemma (Peng and Babcock, 2012;Landing et al., 2013;Babcock et al., 2014;Zhu et al., 2019;Steiner et al., 2020;Yang and Steiner, 2021) very little progress has been made.Reluctance to implement any formal changes may be related to uncertainties and potential limitations concerning alternative marker options.Regardless, it is clear that there are currently two other potential marker options that can assist with the definition of the base of the Cambrian globally; 1) the FAD of the skeletal taxon Protohertzina and 2) the BACE.

The FAD of Protohertzina
The rapid diversification of biomineralizing organisms is a hallmark of the Cambrian radiation and the FAD of a skeletal taxon has long been discussed as a viable option for defining the base of the Cambrian (Cowie, 1981).However, the recent discovery of mixed Ediacaran-Cambrian faunas appears to have influenced many to move away from skeletal taxa as a correlation method (Landing et al., 2013;Zhu et al., 2019).This 'mixing' though is restricted to only a handful of tubular taxa (e.g., cloudiniids, Anabarites and Cambrotubulus), and the suggestion that this faunal overlap indicates an evolutionary continuum between Ediacaran and Cambrian faunas is premature (Zhu et al., 2017;Cai et al., 2019;Bowyer et al., 2022).Many skeletal fossils considered typically Cambrian (based on our current boundary definition) remain exclusively so, such as Protohertzina anabarica.Occurrence of tubular organisms (e.g., Cloudina) in the uppermost Ediacaran (Grant, 1990), is well known, and the presence of 'similar' tubular taxa over the E-C boundary does not diminish the capacity for specific skeletal taxa to define the base of the Cambrian.
Despite the demonstrable correlation potential of Cambrian skeletal taxa, Landing et al. (2013) warned against using the FAD of any fossil to define a global chronostratigraphic unit, due to the diachronous nature of fossil FADs.It is true that speciation is not instantaneous, and there are time lags between appearances of a particular species as it disperses around the globe.Estimates of the global FAD based on local or regional ranges of taxa will tend to underestimate global ranges (Cody et al., 2008;Landing et al., 2013;Smith et al., 2015) and the true stratigraphic distribution of a specific taxon will most likely never be known (Sadler, 2004;Landing et al., 2013;Smith et al., 2015;Lucas, 2018Lucas, , 2019)).In an ideal scenario for the definition of a stage boundary, the FO and the FAD of a taxon will align perfectly, however this is extremely unlikely.In an effort to overcome such problems, Landing et al. (2013) suggested that, rather than the FAD of T. pedum, the GSSP horizon should coincide with the base of the T. pedum Ichnozone Assemblage.The stratigraphic position of the T. pedum Ichnofossil Assemblage Zone is ambiguously defined as "immediately above the last occurrences of H. podolica and P. delicatus" (Landing et al., 2013 p. 145).Regardless, this merely sidesteps the problem as last appearance data (LADs) of fossil taxa share many of the same diachronous concerns as FADs (Sadler, 2004;Cody et al., 2008;Smith et al., 2015).
Stratigraphic concerns over fossil FADs and LADs can be minimized by incorporating stratigraphic records from multiple locations around the world.Subsequent marriage with alternative correlation techniques such as chemostratigraphy or radiometric dates will produce the best estimates of a global FAD (or LAD) for a particular taxon.Protohertzina for example, has been reported from early Cambrian carbonate successions all over the world and its stratigraphic distribution is well documented (e.g., Steiner et al., 2007;Kouchinsky et al., 2012Kouchinsky et al., , 2017)).These occurrences are frequently supported by chemostratigraphic data showing that the FO of Protohertzina is regularly in close association with the prominent negative excursion known as the BACE, and is always above the nadir of this excursion (Fig. 21).This association of the BACE and the first occurrence of Protohertzina across palaeocontinents provides strong evidence to support a near-synchronous distribution (and hence a global FAD) and highlights the taxon's potential for defining the base of the Cambrian.

The nadir of the BACE
δ 13 C profiles are important chronostratigraphic tools and excursions have been used to refine stratigraphy in the Ordovician (Kaljo et al., 2003), Silurian (Cramer et al., 2010;Cramer et al., 2011) and even the Cambrian, where the DICE δ 13 C excursion is a secondary marker of the Drumian GSSP (Babcock et al., 2007).The BACE is recorded in the majority of E-C carbonate successions where similar δ 13 C trends are repeatedly detected.However, despite a concerted push for the BACE to contribute to the definition of the base of the Cambrian (see Zhu et al., 2006Zhu et al., , 2017Zhu et al., , 2019) ) reluctance remains, primarily due to apprehensions regarding the accuracy and reliability of the δ 13 C signal (see Steiner et al., 2020;Yang and Steiner, 2021).Uncertainites affecting the interpretation of δ 13 C chemostratigraphic signals can however be minimized by quantifying local depositional conditions and diagenesis.Identification and correlation of δ 13 C excursions also relies on supplementary biostratigraphic or radiometric markers.The early Cambrian is characterized by several major rapid changes in isotopic ratios, with four negative and five positive excursions in the Fortunian alone (Landing  and Hagadorn, 2000;Corsetti and Hagadorn, 2003;Loyd et al., 2012.Mexico section compiled from data obtained from Sour-Tovar et al., 2007;Loyd et al., 2012Loyd et al., , 2013;;Hodgin et al., 2021.Northwest Canada section compiled from data obtained from Narbonne et al., 1994, Narbonne and Aitken, 1995. Pyle et al., 2006. See Fig. 16 for symbols. et al., 2013;Landing and Kouchinsky, 2016;Kouchinsky et al., 2017).δ 13 C perturbations can be easily confused when not coupled with another proxy (see anomaly 'W' and the BACE; Brasier et al., 1996b;Smith et al., 2016a;Bowyer et al., 2022;herein).Integration of multi-proxy datasets is the most reliable method for interpreting δ 13 C curves.However, while this integrative approach is increasingly utilized, it is yet to become routine.
More commonly, biostratigraphic and chemostratigraphic data are decoupled; collected from several sections at various localities, and often published at different times (e.g., Brasier et al., 1993;Narbonne et al., 1994;Knoll et al., 1995;Smith et al., 2016a).For example, the δ 13 C profile produced by Narbonne et al. (1994) for E-C strata in northwestern Canada was pieced together using data from six localities, and fossil horizons (skeletal and trace fossils) were estimated using previous publications by others (e.g., Conway Morris and Fritz, 1980;MacNaughton and Narbonne, 1992).Similarly, knowledge of the E-C boundary of the Kazakh terranes is pieced together from three disparate localities (Yang et al., 2016;Stammeier et al., 2019;Fig. 18).And until the recent work by Steiner et al. (2020) in Hubei, South China, E-C biostratigraphic data from this area were most commonly collected from outcrop (Guo et al., 2014;Topper et al., 2019), while chemostratigraphic data were obtained from drill core (Ishikawa et al., 2008).
Construction of composite columns is commonly achieved by lithostratigraphic correlation.However, precise characters and thicknesses of lithostratigraphic units will vary across a basin, confounding correlation of a fossil bearing horizon in one section with a chemostratigraphic excursion in another.Therefore, 'stitching' together bio-and chemostratigraphic data from different localities is clearly not best practice, as this obscures the stratigraphic relationship between fossil occurrences and perturbations in the δ 13 C profile.In the Zavkhan Basin for example, incorporation of disparate sections into composite columns, and integration of recent and outdated bio-and chemostratigraphic data has resulted in considerable chronostratigraphic uncertainty (Khomentovsky and Gibsher, 1996;Brasier et al., 1996b;Maloof et al., 2010a;Smith et al., 2016a).Hence, it is clear that fully integrated, multi-proxy chronostratigraphic methods using data derived (preferably simulatenously) from single, continuous sections delivers the most robust and reliable results.Babcock et al. (2014) presented five options for potentially managing the definition of the base of the Cambrian.They include 1) No change in definition, 2) Designate an Auxiliary Stratotype Section and Point (ASSP) to correlate from, 3) Use the best estimate of the FAD of T. pedum, 4) Change the GSSP definition, horizon and evolutionary marker event and, 5) Use a non-biostratigraphic tool to coincide with a redefined GSSP horizon.It is clear that the Cambrian GSSP as currently defined is insufficient for accurate global correlation.Hence, option 1 of Babcock et al. ( 2014) may be eliminated.Option 3 may be practical if the GSSP is redefined at the FAD of T. pedum in another locality, however the exclusive use of T. pedum has been shown to be ultimately imprecise and unreliable.Therefore, either designating an ASSP to enhance correlation of the Fortune Head section or redefining the GSSP entirely (either using a biostratigraphic or non-biostatigraphic marker) represents the most viable options for resolving global correlation of the E-C boundary.

Designating an ASSP
ASSPs represent boundaries that complement the GSSP by using additional stratigraphic markers in a different paleogeographic region to where the GSSP has been defined (Ergaliev et al., 2014).While uncommon, ASSPs have been suggested or designated for a handful of lower Palaeozoic boundaries to promote correlation.For example, Fig. 18.Stratigraphy and chemostratigraphy of Ediacaran-Cambrian successions in northern Gondwana and the Central Asian terranes with occurrences of key ichnotaxa and skeletal fossils.Oman section compiled from data obtained from Amthor et al., 2003;Bowring et al., 2007.Iran section compiled from data obtained from Hamdi et al., 1989;Kimura et al., 1997;Shahkarami et al., 2017.Kazakhstan section compiled from data obtained from Yang et al., 2016;Stammeier et al., 2019.Mongolian section compiled from data presented here.See Fig. 16 for symbols.Fig. 19.Stratigraphy and chemostratigraphy of Ediacaran-Cambrian successions in Siberia with occurrences of key ichnotaxa and skeletal fossils.Dvortsy/Mt Konus section compiled with data obtained from Khomentovsky and Karlova, 1993;Braiser et al., 1993;Kouchinsky et al., 2007Kouchinsky et al., , 2012. .Olenek Uplift section compiled with data obtained from Knoll et al., 1995;Nagovitsin et al., 2015;Rogov et al., 2015.Kotuj River/Kotujkan section compiled with data obtained from Khomentovsky andKarlova, 1992, 1993;Kaufman et al., 1996;Kouchinsky et al., 2007, 2012. See Fig. 16 for symbols.

Fig. 20.
Stratigraphy and chemostratigraphy of Ediacaran-Cambrian successions in South China with occurrences of key ichnotaxa and skeletal fossils.Gunziao section (also known as the Yanjiahe section) compiled from data obtained from Guo et al., 2014;Okada et al., 2014;Steiner et al., 2020.Xiaotan section compiled from data obtained from Zhou et al., 1997;Li and Xiao, 2004;Li et al., 2013;Yang et al., 2016.Laolin section compiled from data obtained from Luo et al., 1991;Li et al., 2009;Yang et al., 2014a.Meishucun section compiled from data obtained from Luo et al., 1984;Zhu et al., 2001;Yang et al., 2014a;Zhang et al., 2020.Markers 'A' and 'B' in the Meishucun section refer to the placement of the E-C boundary in previous studies (see Luo et al., 1984;Zhang et al., 2020).See Fig. 16 for symbols.ASSPs have been suggested to assist with correlating the base of the Ordovician (Miller, 2016;Wang et al., 2021) and an ASSP was established in Kazakhstan for the base of the Cambrian Jiangshanian Stage (Ergaliev et al., 2014).A pathway forward for the base of the Cambrian may be to designate an ASSP at a locality where both T. pedum and an additional marker (such as the BACE) occurs.Stratigraphic sections in Laurentia (e.g., Boundary Canyon and Mount Dunfee sections in California and Nevada and the Cerro Rajón section in Mexico; Corsetti and Hagadorn, 2000;Smith et al., 2016b;Hodgin et al., 2021) preserve both markers and a radioisotopic age (the Cerro Rajón section) and therefore have potential for defining an ASSP.Several disadvantages of the current GSSP could be improved by the designation of an ASSP and the incorporation of a secondary marker such as the BACE.However currently, the establishment of a Cambrian ASSP is undermined by the total lack of additional secondary markers at Fortune Head.As the stratigraphic relationship between T. pedum and secondary markers is unknown in the GSSP section, reliable correlation between the current GSSP and a potential ASSP would not be possible.

Redefining the GSSP
Complications involved with establishing an ASSP highlights the need to reevaluate the definition, horizon and marker of the current Cambrian GSSP.Fears that redefining the GSSP would involve the 'destabilization of the definition of the Cambrian System' (Babcock et al., 2014, p. 9) have discouraged progress.However, it is clear that the base of the Cambrian is already unsettled, and long-term term consistency should be favoured over short-term destabilization.Reassessment of the Cambrian GSSP is not without precedent.In 2008 the GSSP for the base of the Silurian system was redefined following confusion over the biostratigraphic definition of the boundary (Rong et al., 2008).This has ensured a greater understanding of the boundary interval and improved the resolution with which it can be globally correlated.Skeletal fossil biostratigraphy (Steiner et al., 2020;Yang and Steiner, 2021), carbon isotope chemostratigraphy (Zhu et al., 2017(Zhu et al., , 2019) ) and ichnofossil occurrences (Landing et al., 2013;Buatois, 2018) are all possible alternative methods to define the base of the Cambrian.However, all are affected to some degree by preservational biases or diagenetic overprinting.Hence the capacity of each marker to independently define a global boundary is doubtful.
The primary purpose of a GSSP is to promote correlation over a large area and reliance on a single marker for robust correlation is unacceptable.Secondary proxies are essential for permitting correlation with localities where the primary signal is absent (Finney, 2013;Babcock et al., 2014;Lucas, 2018Lucas, , 2019)).This multi-proxy chronostratigraphic approach is now a requirement of the International Commission on Stratigraphy (ICS) for the definition of chronostratigraphic units (Cowie et al., 1986;Remane et al., 1986;Babcock et al., 2014).Yet, despite this emphasis on adopting a multi-proxy approach, the notion that a single marker may be sufficient for defining the E-C boundary persists.For example, Zhu et al. (2019, p. 36) state that "the base of the Cambrian can be redefined as the onset of decreasing δ 13 C values from the terminal Ediacaran positive carbon isotope plateau to the BACE excursion" and do not suggest secondary markers.
Faunal assemblages across the E-C boundary comprise only a handful of skeletal fossils, acritarchs and ichnotaxa, offering much less diversity than at younger boundaries.The Albian Stage of the Cretaceous for example has 28 secondary markers including calcareous nannofossils, planktonic foraminifera, ammonites and a bivalve (Kennedy et al., 2017).Diverse secondary taxa provide ample opportunities to correlate, and mitigate the effects of strong facies dependence of a single marker.However, this approach may only be applied in the Cambrian if the boundary is moved significantly higher in the geological timescale (to the base of Stage 3 at the incoming of trilobites for example).Currently, only three proxies have strong potential for globally correlating basal Cambrian strata; the FAD of T. pedum, the FAD of P. anabarica and the BACE.While it is clear that the GSSP horizon at Fortune Head does not represent the best option to define the base of the Cambrian globally, dismissing T. pedum as a stratigraphic marker entirely (Zhu et al., 2017(Zhu et al., , 2019;;Steiner et al., 2020;Yang and Steiner, 2021) further reduces opportunities for correlation, and limits definition of the boundary to even fewer markers, decreasing capacity for reliable global correlation.
In the few regions where all three markers have been unambiguously documented, they occur in relatively close association, with the FO of T. pedum and P. anabarica both occurring stratigraphically above the nadir of the BACE (Figs. 17-21).In the Mackenzie Mountains in Canada, the FO of T. pedum and P. anabarica and the nadir of the BACE all occur within the Ingta Formation (Narbonne et al., 1994;Figs. 17, 21).In South China, the nadir of the BACE is in the Daibu Member and the FO of T. pedum and P. anabarica occur in the overlying Zhongyicun Member (Li et al., 2009(Li et al., , 2013;;Figs 20,21).Unfortunately, in these regions all three markers are yet to be documented from the same stratigraphic section.There are however, several locations around the globe where two out of the three markers have been documented from the same section.
The stratigraphic relationship between the BACE and the FAD of T. pedum and the FAD of P. anabarica are consistent globally, and represent perhaps the most promising way to redefine the base of the Cambrian.Treptichnus pedum and the BACE have been documented from a number of sections in Laurentia (Fig. 21), including the Boundary Canyon section in Death Valley (Corsetti and Hagadorn, 2000), the Cerro Rajón section in Mexico (although the markers are separated by an unconformable surface; Hodgin et al., 2021) and at Mt Dunfee in Esmeralda County (most likely section E1421 of Smith et al., 2016a).Protohertzina anabarica and the BACE have been reported from the Laolin (Li et al., 2009), Xiaotan (Li et al., 2013) and Gunziao (Steiner et al., 2020) sections in South China, the Kotuj River/Kotujakan section 2 (Kaufman et al., 1996;Kouchinsky et al., 2017) in Siberia and the Bayan Gol section in Mongolia documented herein.
In the absence of a single stratigraphic section that yields all three markers we propose that a new GSSP and corresponding ASSP be erected; one section that captures the FO of T. pedum and the BACE, and another that captures the FO of P. anabarica and the BACE.The nadir of the BACE would correlate the two stratotype sections.δ 13 C chemostratigraphy represents the best solution to enhance correlation across facies and palaeocontinents.However, because identification of the BACE requires support from biostratigraphic data, both P. anabarica and T. pedum would serve as essential secondary markers.In this scenario, in siliciclastic successions where a δ 13 C profile is unavailable, the FAD of T. pedum may be implemented.In carbonate successions, either the BACE or the FAD of P. anabarica may be employed as the boundary marker.Erection of both a new GSSP and an ASSP gleans as much utility as possible from the few chronostratigraphic markers that are available in the early Cambrian.
We suggest that the GSSP horizon be defined in a predominantly carbonate succession, as this would provide the most complete δ 13 C profile.Because samples for isotopes and skeletal fossils can be collected simultaneously, a carbonate section also provides the opportunity to record the FAD of the secondary marker P. anabarica.Other skeletal fossils (such as cloudiniids or Anabarites) commonly found at these stratigraphic levels would provide additional biostratigraphic data to bracket the BACE.Should the Cambrian GSSP be redefined, the Bayan Gol section in Mongolia (documented herein) represents a strong candidate, as do sections in South China (Xiaotan; Fig. 20) and Siberia (the Kotuj River/Kotujakan section 2; Fig. 19).Laurentian sections from the Great Basin that preserve the FAD of T. pedum in close association with the BACE (Fig. 17) are likely candidates for a Cambrian ASSP.

Conclusion
The application of T. pedum has demonstrated that a single proxy is insufficient to define the base of the Cambrian.Lack of alternative markers (e.g., skeletal fossils, chemostratigraphic, magnetostratigraphic or radiometric data) at the GSSP locality at Fortune Head isolates this section from global correlation, and has prompted the search for different ways to define the boundary (Babcock et al., 2014;Zhu et al., 2019).There is now a general consensus that the integration of multiple proxies (such as biostratigraphy and chemostratigraphy) should be employed for global correlation and boundary definition (Betts et al., 2018;Zhu et al., 2019;Steiner et al., 2020).For the base of the Cambrian, three potential markers may be used; the FAD of T. pedum, the FAD of P. anabarica and the nadir of the BACE.The nadir of the BACE represents the most promising way to redefine the base of the Cambrian, but employing these three proxies in conjunction strengthens their application towards global correlation of the E-C boundary.
Bayan Gol in western Mongolia represents a unique opportunity to document shelly fossils, isotopes and trace fossils through a continuous, predominantly carbonate succession that captures the E-C transition.Although this section has been studied extensively in the past (Voronin et al., 1982;Brasier et al., 1996a;Khomentovsky and Gibsher, 1996;Esakova and Zhegallo, 1996;Smith et al., 2016a), the practice of developing composite sections and decoupling fossil and isotope data has obscured the stratigraphic significance of this sedimentary package.Shelly fossil biostratigraphy with tightly integrated δ 13 C data demonstrates the close temporal relationship between the FAD of Protohertzina and the nadir of the BACE in the upper Zuun-Arts Formation at Bayan Gol.This stratigraphic relationship is consistent globally, making the section at Bayan Gol an ideal candidate for a redefined Cambrian GSSP.Redefining the GSSP and defining a corresponding ASSP by exploiting the stratigraphic relationships between the BACE and other accessory markers is the most robust method of globally correlating the E-C boundary.
GSSPs are often regarded as immutable points (Ager, 1963;Cowie et al., 1986;Vai, 2001), that once ratified should "remain fixed in spite of discoveries stratigraphically above and/or below" (Cowie, 1986, p. 79).However, new discoveries and advances in methods and techniques will undoubtedly fine-tune these boundaries.Biostratigraphic definitions have historically dominated stage boundaries through the Phanerozoic (Smith et al., 2015), but δ 13 C trends are increasingly utilized for global correlation.Steadfast refusal to incorporate new data that improves chronostratigraphic boundaries prevents robust global correlation of the Cambrian and undermines scientific progress.The Ediacaran-Cambrian boundary is one of the most significant boundaries in the geological timescale.The boundary separates the Proterozoic and the Phanerozoic eons, marking major revolutions in the Earth's geosphere and biosphere.Ability to reliably identify and correlate this boundary globally is fundamental for charting evolutionary dynamics across this transition.Utilizing all available markers to redefine the GSSP and define a corresponding ASSP represents a pathway toward global correlation of the base of the Cambrian, and an essential improvement on the GSSP at Fortune Head.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 2 .
Fig. 2. Interpretations of the lithostratigraphy of Bayan Gol as documented byKhomentovsky and Gibsher (1996) and our interpretation herein.A, Stratigraphic log based on the geological mapping ofKhomentovsky and Gibsher, 1996, fig.6), note the repetition of units 17 and 18 of the Zuun-Arts and Bayangol formations.B, Lithostratigraphic log presented byKhomentovsky and Gibsher, 1996, fig.13)  where the interpreted faulted strata has been omitted resulting in a condensed log and the rearrangement of sample horizons (eg.RIII-X).C, our interpretation that the lower Bayangol Formation represents continuous stratigraphy, resulting in a much thicker sequence and the repositioning of samples RIV-RX.

Fig. 3 .
Fig. 3. Lithostratigraphic column for the BAY1 section through the lower Zuun-Arts Formation in Bayan Gol, southwestern Mongolia, including ranges of fossil taxa and δ 13 C and δ 18 O chemostratigraphic data.Samples taken are listed on the lithostratigraphic column.Sh=Shuurgat Formation.

Fig. 5 .
Fig. 5. Lithostratigraphic columns for the BAY4 and BAY5 sections through the upper Zuun-Arts and lower Bayangol formations in Bayan Gol, southwestern Mongolia, including ranges of fossil taxa and δ 13 C and δ 18 O chemostratigraphic data.Note: BAY4/13.0 = BAY5/0.0.Samples taken are listed on the lithostratigraphic column, red samples are from the BAY4 section.PC=Precambrian.

Fig. 11 .
Fig. 11.Thin section photomicrographs of fabrics and textures from the lower Bayangol Formation in the BAY5 section, Bayan Gol.A, BAY5/151.1,dolomitisation front (lower part of image) against recrystallized micritic limestone with abundant dissolution features.B, BAY5/177, mottled micrite with abundant recrystallisation to calcite spar and fine, thread-like dissolution veins.C-H, BAY5/196.5,fine, reworked intraclasts in a light micritic matrix.Intraclasts are sub-rounded and are a variety of shapes with stippled interiors.D-H, intraclasts from BAY5/196.5 may be calcareous metazoan fossil or microbial limestone fragments.They show stippling (dark in places likely due to limonisation or replacement by hematite) that is occasionally aligned (F).I-K, BAY5/197.2,grey, mottled micrite with abundant reworked and randomly oriented intraclasts (J).Intraclasts often recrystallized to calcite spar.Much of the original fabric has been obliterated by pervasive dolomitisation (K).L-N, BAY5/222.8,light, grainy micrite with "shadows" of elongate (occasionally aligned) intraclasts and pellets.Pellets commonly contain fine, angular quartz grains (L).

Fig. 12 .
Fig. 12. Trace fossils from the lower Bayangol Formation (basal Fortunian) at Bayan Gol, southwestern Mongolia.Both specimens derived from a sandstone layer at BAY5/108.5.A-B, simple Planolites-like traces and wider, tri-lobed Psammichnites-like trace on bed sole NRM X036.C-E, Tri-lobed Psammichnites-like trace.C, bed sole with trace in positive relief, NRM X037a.D, counterpart with trace in negative relief, NRM X037b.E, close-up of C showing trilobed nature of the trace.