1 Introduction
The pre-Variscan succession of the Montagne Noire crops out as a fold-and-thrust complex divided into two sedimentary-dominated, northern and southern flanks fringing an essentially metamorphic, Axial Zone (Arthaud, 1970; Gèze, 1949). Several tectonic models have been proposed for the Montagne Noire and are still subject of debate (Brun and Van Den Driessche, 1994; Charles et al., 2009; Faure and Cottereau, 1988; Fréville et al., 2016; Mattauer et al., 1996; Poujol et al., in press; Soula et al., 2001; Van Den Driessche and Brun, 1992). The Precambrian–Cambrian boundary has traditionally been tentatively located in the lowermost formation exposed in the southern Montagne Noire, namely the Marcory Formation (Álvaro et al., 1998). However, recent reviews of the northern successions challenged the former stratigraphic chart (Álvaro et al., 2014b; Devaere et al., 2013) and established a lower stratigraphic position for the volcano-sedimentary successions of the Grandmont, Rivernous and Layrac formations, exclusively exposed in the northern Montagne Noire. The only previous radiometric data from the Rivernous Formation (Fm.) yielded an Ordovician age, ranging from 473 ± 19 Ma to 443 ± 40 (Rb/Sr method; Demange, 1982), which was subsequently ruled out by biostratigraphic ages yielded by acritarch-bearing, laterally equivalent deposits (the so-called “Schistes X”) from the Axial Zone (Fournier-Vinas and Debat, 1970). There, the “Schistes X” are capped by the Sériès Tuff (“Sériès” is a village name) tentatively correlated with the Rivernous Fm. The Sériès Tuff was dated at 545 ± 15 Ma by Pb-evaporation on zircon from a metadacite (Lescuyer and Cocherie, 1992). Some scarce zircon crystals were sampled in a garnet-grade Cambrian meta-siltstone from the southern Montagne Noire, giving a maximum depositional age of 556 Ma based only on a single zircon (Gebauer et al., 1989). This age attribution may look disputable, given the more recent statistical guidelines for provenance studies (see below) and also taking into account the marked metamorphic character of this rock. In fact, in a previous paper, Gebauer and Grünenfelder (1977) admitted that about 80% of the primitive radiogenic Pb might have been lost due to Phanerozoic thermal events. In order to improve the stratigraphic framework of the Neoproterozoic–Cambrian boundary interval, and the lithostratigraphic nomenclatural links between the Axial Zone and the northern and southern flanks of the Montagne Noire, zircon grains were sampled from the Rivernous, Grandmont and Marcory formations and dated by in situ LA-ICP-MS U–Pb analysis. Our results place new constraints upon the palaeogeographic affinities of the different tectonostratigraphic units that form the Montagne Noire, as well as on the detrital provenance of the Ediacaran–Cambrian sediments preserved in neighbouring tectonostratigraphic areas.
2 Geological and stratigraphic setting
Located in the southern part of the French Massif Central (Fig. 1A), the Montagne Noire represents a segment of the external, southwestern component of the Variscan Belt in Europe (Demange, 1998; Poujol et al., in press; Roger et al., 2004). As summarized above, this ENE–WSW-striking range is divided into three tectonic units: a central metamorphic dome, the so-called Axial Zone, fringed by its northern and southern flanks (Demange, 1985; Gèze, 1949).
Simplified geological map of the Montagne Noire; modified from Devaere et al. (2014). A, Location of the French Massif Central (grey) and Montagne Noire (rectangle) in France. B, Structural units and previous radiometric ages: (a) Demange, 1982 (Rb/Sr) discarded by our results (see text); (b) Ducrot et al., 1979 (U–Pb) in Demange et al., 1995; (c) Lescuyer and Cocherie, 1992 (U–Pb); (d) Roger et al., 2004.
The Axial Zone is essentially composed of micaschist, minor marble, paragneiss and migmatized orthogneiss (Gèze, 1949). The protolith age of the orthogneiss and its relationship with the metasedimentary rocks have been disputed. Some authors interpreted the orthogneiss as a granitic Precambrian basement (Demange, 1975, 1998), whereas others considered it as Palaeozoic intrusions (Bard and Loueyit, 1978). Recent conventional (ID-TIMS), SHRIMP and LA-ICP-MS U–Pb datings of various orthogneiss samples (Cocherie et al., 2005; Pitra et al., 2012; Roger et al., 2004) suggest that the granitic protolith was emplaced during the Ordovician.
The southern and northern sedimentary-dominated flanks of the Axial Zone are a fold-and-thrust complex of nappes (Fig. 1). The Precambrian–Cambrian boundary interval, only reported in the northern Montagne Noire (Álvaro et al., 2014b), comprises four formations, from bottom to top, the Grandmont, Rivernous, Layrac and Marcou formations (Fig. 2). The Grandmont Fm., about 700 m thick, consists of grey to black shales with subsidiary sandstone interbeds (Fig. 2). The Rivernous Fm., up to 200 m thick, comprises slightly metamorphosed rhyolitic tuffs that include rare breccia and shale interbeds (Fig. 2). Both formations crop out in the Avène–Mendic parautochthon, which includes the Lodévois inlier and the Lacaune thrust slice (Murat, Fig. 1B). In the Lacaune unit, a lateral equivalent of the Rivernous Fm. (locally named Murat Fm.), with base and top truncated by faults, was dated at 532 ± 12 Ma (U–Pb on zircon; Demange et al., 1995; Ducrot et al., 1979). In the Avène–Mendic parautochthon (Fig. 1), the rhyolitic palaeorelief formed by the Rivernous Fm. (up to about 300 m high, after Álvaro et al., 2014b) is unconformably onlapped by the volcano sedimentary conglomerates and sandstones of the Layrac Fm. (Fig. 2). The Layrac Fm. is itself overlain by the carbonate-dominated Marcou Fm., about 400 m thick and assigned to the Cambrian Stage 2 by recent biostratigraphic studies (Fig. 2, Devaere et al., 2013).
Upper Ediacaran–lower Cambrian stratigraphic chart of the southern and northern Montagne Noire; modified from Álvaro et al. (2014b).
The above-reported formations are not exposed in the southern Montagne Noire, where the oldest outcrop is represented by the up to 1000 m-thick Marcory Fm. (Fig. 2), a monotonous alternation of sandstones and shales with subordinate carbonate nodules and layers. The upper part of the Marcory Fm. has been assigned to the Cambrian Stage 2–3 transition due to the occurrence of the ichnogenera Psammichnites and Taphrelmintopsis (Álvaro and Vizcaïno, 1999), and the oldest trilobites found in the Montagne Noire, i.e. Blayacina miqueli (Cobbold, 1935; Geyer, 1992). The Marcory Fm., although absent in the Avène–Mendic parautochthon, is exposed in other thrust slices and nappes of the northern Montagne Noire.
3 Material and methods
3.1 Material
The Grandmont and Rivernous Formations have been sampled at their stratotype, along the Rivernous rivulet (about 4 km to the east of Lodève): a sandstone from the base of the exposed Grandmont Fm. (sample MN1), and a rhyolitic tuff from the base of the Rivernous Fm. (sample MN2). The Rivernous Fm. was also sampled at its middle part near the Col du Layrac (sample MN3). MN1 to MN3 samples belong to the so-called Avène–Mendic parautochthon. In order to investigate the provenance of the Marcory Fm., a sandstone (sample MN4) was also selected from the Psammichnites gigas-bearing level in the southern Montagne Noire (Álvaro and Vizcaïno, 1999), along the Orbiel river section of the Minervois nappe (Fig. 1).
3.2 Samples preparation
Zircon separation from fresh samples started with rock grinding using a steel crusher. The resulting powders were sieved in the range of 50–250 μm. Grains were separated first using a heavy liquid (sodium heteropolytungstates, density 2.85 g·cm–3), then using a Frantz magnetic separator. Following Sláma and Košler (2012), the selected grains were obtained from random handpicking under a binocular microscope whatever their size, shape, or color, in order to avoid any operator bias. They were finally set in an epoxy resin puck and polished to expose their core.
3.3 LA-ICP-MS in situ U–Pb dating
To identify internal growth textures and morphologies, zircon grains were imaged by scanning electron microscopy (SEM) to get cathodoluminescence and back-scattered electron images (at the “Laboratoire d’océanologie et de géosciences”, University of Lille, France). The U–Pb ages of zircons were determined in situ at the Géosciences Rennes laboratory by LA-ICP-MS using an ICP-MS Agilent 7700x coupled with an ESI laser Excimer system producing a radiation with a wavelength of 193 nm (NWR193UC), with ablation spot diameters of 25 μm, energy pulses of 7J·cm–2, and repetition rates of 5 Hz. Ablations were operated on both grain rims and cores. Where necessary, distinct domains of a zircon grain were analyzed to compare their ages. The resulting ablated material was mixed in a He, N and Ar gas mixture before being transferred into the plasma source of the ICP-MS device. Each analysis lasted 80 s and consisted of a first ∼20-s background measurement followed by ∼60-s ablation with measurements of 204(Hg + Pb), 206Pb, 207Pb, 208Pb, 232Th, and 238U, and a ∼15 s wash-out delay before the next acquisition. The data were collected in batch of 43 analyses divided in three sets of 10 unknowns, bracketed by two measurements of the GJ-1 primary zircon standard (Jackson et al., 2004) to correct for U–Pb and Th–Pb laser-induced fractionation and for instrumental mass discrimination, followed by one analysis of the Plesovice secondary zircon standard (Sláma et al., 2008) in order to check the precision and accuracy of the measurements. During the course of this study, the Plesovice zircon standard yielded a Concordia age of 336.8 ± 0.67 Ma (N = 32). The operating conditions for the LA-ICP-MS equipment can be found in Supplementary Table 1. For more information on the acquisition protocol, see Manzotti et al. (2015). Data treatment was performed with the GLITTER software (Van Achterbergh et al., 2001) and plotted using the Isoplot 3.75 software (Ludwig, 2012) in both Wetherill and Tera-Wasserburg Concordia diagrams. For rhyolitic tuffs, the ages were calculated using the TuffZirc Age algorithm (Ludwig and Mundil, 2002) together with the Sambridge and Compston (1994) algorithm. For the sandstones, age distribution curves with probability density plot were obtained using the density plotter freeware proposed by Vermeesch (2004). For dates > 1 Ga, we reported the 207Pb/ 206Pb dates and for ages < 1 Ga, we used the 206Pb/238U dates. The analyses out of the [90–110%] concordance interval, calculated with 100 × (207Pb/235U age)/(207Pb/206Pb age) for ages > 1 Ga (Meinhold et al., 2011) and 100 × (206Pb/238U age)/(207Pb/235U age) for ages < 1 Ga, were rejected (Faure and Mensing, 2005 and Talavera et al., 2012). The age of the youngest zircon population is derived from a cluster of at least three analyses from three different grains overlapping in age at 2σ (standard deviation), as proposed by Dickinson and Gehrels (2009) to ensure a statistically robust estimate of the maximum depositional ages. Percentages of concordance, isotopic ratios and ages with 1σ errors, as well as U and Pb concentrations are provided in Supplementary Table 2. In sedimentary rock samples (MN1 and MN4), about 110 grains were analyzed in order to get the best representation of the detrital zircon populations. For tuffs (MN2 and MN3), about 50 grains were analyzed, following the suggestions of Bowring et al. (2006), to give a robust estimate of the best age for the related volcanic event(s).
4 Results
4.1 Grandmont Formation (MN1)
Zircons from the sample MN1, medium-grained sandstone, are mostly in the 100–250 μm range, euhedral, facetted, rarely sub-rounded, colourless and generally well zoned (Supplementary data, Fig. S1). 107 of the 114 MN1 analyses were concordant [90–110%], among which 94% are Neoproterozoic (101 grains), 3% Mesoproterozoic (3 grains), 2% Paleoproterozoic (2 grains) and 1 grain (1%) is Archean in age (Fig. 4B). In the interval ranging from 500 to 1100 Ma, the probability curve shows a dominant Ediacaran group within the cluster 550–850 Ma, which displays a main peak around 605 Ma and a secondary one around 635 Ma (Fig. 4B). In this same cluster, four subsidiary peaks are identified around 690 Ma, 760 Ma, 805 Ma and 835 Ma. The Tonian-aged zircon grains are characterized by one peak around 906 Ma, in an 890–920 Ma cluster and another peak around 1 Ga in the 950–1050 Ma cluster. The four youngest and concordant zircon grains from this group ranging from 567.2 ± 6.12 Ma to 579.5 ± 6.28 Ma, yield an average date of 574 ± 6 Ma (95% conf., Fig. 3E).
Frequency and probability density plots of detrital zircon grains in the range 500–3300 Ma for samples MN1 (B) and MN4 (A). Age groups of each sample are presented in a pie diagram. Section C shows a comparison of age groups where white boxes represent dates from inherited core and black boxes relate to dates from zoned rims.
LA-ICP-MS results for all samples using 206Pb/238U ages. Diagram A: mixture analysis for sample MN2 using Sambridge and Compston's (1994) approach; the age 579.8 ± 4.3 Ma is considered as inherited age. Diagram B: mixture analysis for sample MN3 using Sambridge and Compston (1994) approach; the age 524 ± 7.2 Ma is considered as geologically meaningless, due to lead loss. Diagrams C and D: data point age distribution for samples MN2 and MN3, resp., using TuffZirc Age (Ludwig and Mundil, 2002); chosen emplacement age for the rhyolitic tuffs of the Rivernous Fm. are: 542 ± 3 Ma (2σ) for sample MN2 and 537 ± 3 Ma (2σ) for sample MN3. Diagrams E and F: average age calculated for the youngest concordant zircon population of Grandmont and Marcory Fm. (sample MN1 and MN4), derived from a cluster of at least three analyses from three different grains overlapping in age at 2σ (standard deviation) as proposed by Dickinson and Gehrels (2009).
4.2 Rivernous Formation (MN2 and MN3)
Samples MN2 and MN3 are both rhyolitic tuffs. Accordingly, zircon grains from samples MN2 and MN3 are euhedral, clear and colourless. Rarely cored, internal structures highlighted by the CL-imaging show clear magmatic zoning (Supplementary data, Fig. S1).
In sample MN2, 59 zircon grains were analysed, among which 50 gave concordant dates [90–110% as defined above]. Seven of the 50 analyses were rejected due to possible lead loss. Thus the youngest and main population from this sample represent 65% of these 43 zircon grains with individual 206Pb/238U dates ranging from 534.8 ± 6.01 Ma to 546 ± 6.02 Ma. The second group represents 23% of the total population, and gives 206Pb/238U dates ranging from 576.9 ± 6.3 Ma to 600.8 ± 6.7 Ma (Fig. 3A). Finally, five inherited core grains were dated at 651.6 ± 7.2 Ma, 661.8 ± 7.1 Ma, 662 ± 37.1 Ma, 685.8 ± 7.6 Ma and 704.7 ± 7.6 Ma (Supplementary Table 2).
Looking at the youngest population, the TuffZirc Age algorithm returned a date of 542.5 +0.7/–0.6 Ma, while the Sambridge and Compston algorithm yielded a comparable date of 541.9 ± 2.3 Ma (Fig. 3A and C). Therefore, choosing between those two within error identical results, we conclude that this rhyolitic tuff was emplaced 542.5 +0.7/–0.6 Ma.
In sample MN3, 58 zircon grains were analysed, among which 52 yielded concordant results [90–110%]: two composite grains with cores (Zr 1 and Zr 27) yielding U–Pb dates of 1837.7 ± 19.1 Ma and 583.1 ± 6.37 Ma respectively. The youngest zircon population suggested by the Sambridge and Compston approach for sample MN3 returned a date around 524 Ma (Fig. 3B), which is poorly constrained, by only one concordant zircon among seven somewhat discordant data points (on a Concordia plot). Therefore, this date is interpreted as geologically meaningless as it could be linked to a possible lead loss.
Keeping only a group comprising three concordant data points (grey bars in Fig. 3B), the TuffZirc Age algorithm yielded a date of 537.35 +2.35/–1.25 Ma while the Sambridge and Compston approach returned a comparable date of 537.1 ± 2.5 Ma. This second rhyolitic sample was therefore emplaced 537.1 ± 2.5 Ma ago (2σ) (Fig. 3B and D).
4.3 Marcory Formation (MN4)
Sample MN4 is a fine-grained and mature sandstone. Accordingly, its zircon crystals are mostly anhedral, rounded to subrounded, often broken and rarely bi-pyramidal, largely in the 50–100-μm size range. They are colourless to yellowish, though the biggest zircon grains are reddish in colour. In the analysed fraction, 104 of the 112 MN4 analyses were concordant [90–110%], among which 87% are Neoproterozoic (90 grains), 4% Mesoproterozoic (4 grains), 5% Paleoproterozoic (6 grains), 2% Neoarchean (2 grains), one grain is Mesoarchean and the oldest one is Paleoarchean in age at 3.2 Ga (Fig. 4A).
The probability density curve (Vermeesch, 2004) shows a dominant Ediacaran group (clustered across 550–850 Ma), with a main peak around 614 Ma and a secondary peak around 575 Ma (Fig. 4A). In this same cluster, six other peaks are identified around 651 Ma, 678 Ma, 700 Ma, 737 Ma, 800 Ma and 850 Ma. The Tonian-aged zircon grains are characterized by one peak around 900 Ma in an 890–920 Ma cluster and another distinctive peak around 1 Ga in the 950–1050 Ma cluster. The three youngest dates obtained from this sample that are concordant yield an average date of 602.5 ± 7.3 Ma (Fig. 3F).
5 Discussion
5.1 The Rivernous volcanic activity marking the Precambrian–Cambrian boundary interval
The two dates (537.1 ± 2.5 Ma and 542.5 +0.7/–0.6 Ma) obtained from the Rivernous rhyolitic tuffs are much older than previously estimated (473 ± 19 Ma and 443 ± 40 Ma; Demange, 1982). These results allow us to confidently identify the Precambrian–Cambrian boundary (541 Ma, Gradstein et al., 2012) in the basal succession of the Montagne Noire. By comparison with the previously determined age of the Sériès Tuffs from the Axial Zone (545 ± 15 Ma; Lescuyer and Cocherie, 1992) and the Rivernous (former Murat) Fm. in the Lacaune thrust slice of the northern flank (532 ± 12 Ma, intercept superior; Demange et al., 1995 and references therein; Figs. 1–2), these results support the lateral equivalence of the Sériès volcanic episode (Axial Zone) and Rivernous rhyolitic tuff deposition (northern flank), as already suggested (Pouclet et al., in press).
In other Variscan units of the Ibero-Armorican Arc, some plutonic bodies have recently been dated around 540 Ma by LA-ICPMS (Supplementary data, Fig. S2; see also Casas et al., 2015; Castiñeiras et al., 2008; Gutiérrez-Alonso et al., 2004; Melleton et al., 2010; Rubio-Ordóñez et al., 2015): the Arc-de-Fix and Ardéchois augen gneisses in the Massif Central, with respective Concordia age of 541.8 ± 3.1 and 542.5 ± 3.1 Ma (Couzinié et al., this issue), as well as the Laparan orthogneiss in the Central Pyrenees, with a Concordia age of 545 ± 3 Ma (Mezger and Gerdes, 2016). All these plutonic and volcanic events, overlapping in age within error, should be linked to a common episode associated with the voluminous magmatic and anatectic Cadomian events reported for West Gondwana, among others, by Linnemann et al. (2007, 2008). According to these authors, the numerous plutonic and volcanic to volcano sedimentary complexes identified in the Ossa-Morena, Saxo-Thuringian and Anti-Atlas zones (Álvaro et al., 2014a; Blein et al., 2014) can result from a slab breakoff of a southward subducted oceanic plate ending with the Cadomian cycle at about 545–540 Ma. The end of the Pan-African/Cadomian cycle led to the onset of a Cambrian magmatic cycle (Supplementary data, Fig. S2), related to the rifting of the North Gondwana margin (Álvaro et al., 2014a, 2014b; Pouclet et al., in press).
5.2 Age and potential provenances of the Grandmont and Marcory formations
Depositional ages. The youngest group of concordant zircon grains from the Grandmont Fm. (sample MN1) yield an average date of 574 ± 6 Ma that is interpreted here as its maximum depositional age (i.e. late Ediacaran; Fig. 3E). This result is coherent with the stratigraphic framework proposed by Álvaro et al. (2014b) and the Neoproterozoic age based on acritarchs reported from the “Schistes X” Fm. of the Axial Zone (Fournier-Vinas and Debat, 1970). As a result, this maximum age of deposition supports their correlation between the Grandmont and “Schistes X” formations (Álvaro et al., 2014b).
The average date calculated from the youngest group of concordant data for the Marcory Fm. at 602.5 ± 7.3 Ma (Fig. 3F) might be considered as a maximum depositional age. However, it is far from the real depositional age, as mentioned above. Indeed, this sample was selected from the Psammichnites gigas-bearing level (Fig. 2), and consequently must be assigned to the Cambrian Stage 2–3 transition, i.e. it is less than 521 Ma old.
Source of the pre-ca. 1 Ga detrital zircon grains. The age spectra obtained for the samples from the Marcory (MN4) and Grandmont (MN1) formations are somewhat similar, but display some differences. Although each individual singleton-date should be treated with caution, the distributions of pre-1 Ga ages are plainly distinct (Figs. 4 and 5).
Potential sources for samples MN1 and MN4 (modified after Drost et al., 2011; Linnemann et al., 2011; Pereira et al., 2011; Pereira et al., 2012; Tack et al., 2001; Tack et al., 2010). AC1: Siderian event of the Amazonian craton; AC2: Rhyacian, Orosirian and Statherian events of the Amazonian craton; AC3: San Ignacio and Sunsas events of the Amazonian craton; AC4: eastern margin of the Amazonian craton; Av: Avalonia; Cd: Cadomia; Bo: Bohemian massif; AA: Anti-Atlas; WAC1: Eburnean event of the West African craton; WAC2: Liberian event of the West African craton; WAC3: Leonan event of the West African craton; TSB1: Trans-Saharan belt, Benin-Nigerian shield; TSB2: Trans-Saharan belt, Tuareg shield; SMC: Saharan metacraton; ANS: Arabian–Nubian shield; CC: Congo craton.
As mentioned above, zircon grains from the Marcory Fm. (sample MN4) are much smaller and definitely more rounded (Supplementary data, Fig. S1) than those from the Grandmont Fm. (sample MN1), suggesting a long distance of transport for the former grains. We have separated data from inherited cores (white boxes in Fig. 4), meaningless regarding the source age, from data from zoned magmatic rims (black boxes, Fig. 4) that represent most probably the age of the source-rock. From this respect, the oldest zircon rim from the Grandmont Fm. (MN1) is only Late Paleoproterozoic (1806 ± 19 Ma), and only two grain rims are older than 1 Ga. By contrast, the presence of several Archean zircon rims in MN4 points to a major change in the source areas. MN1 zircon grains could possibly be derived, ultimately, from the Amazonian craton (Rhyacian, Orosirian and Statherian events), more probably from the Eburnean West African craton, or from the Saharan metacraton (Fig. 5). By contrast, eight pre-1Ga zircon rims from sample MN4 form a Paleoproterozoic group (cluster ranging from 2109 ± 19 Ma to 1787 ± 20 Ma) and an Archean group. Those groups fit much better with an African source comprising Eburnean areas plus Archean West African craton (Fig. 5).
However, we should not exclude that the availability of zircon to erosion and transport from either primary crystalline or recycled sources requires that the zircon-bearing rocks be exposed at the appropriate time, and recycling from older sedimentary deposits may constitute a more significant source than from primary crystalline rocks.
Sources of the Stenian–Tonian detrital zircons. The cluster pointing to the Stenian–Tonian transition (950 to 1100 Ma) in samples MN1 and MN4 could share a common origin (Fig. 4). One should expect an Amazonian craton affinity (Linnemann et al., 2011) with detrital sources related to the San Ignacio and Sunsas events and on its eastern margin (Pereira et al., 2012). However, many studies on sandstones from Lower Palaeozoic peri-Gondwanan exposures around the Mediterranean region (e.g., Israel, Jordan, Libya, Pyrenees, Sardinia, Greece and Sicily) rule out an Amazonian provenance and suggest instead an eastern to southeastern African origin (Altumi et al., 2013; Avigad et al., 2003; Avigad et al., 2012; Kydonakis et al., 2014; Margalef et al., in press; Meinhold et al., 2011; Meinhold et al., 2013; Williams et al., 2012). Therefore, the Arabian–Nubian shield, the Saharan metacraton, possibly the western edge of the Congo craton (Tack et al., 2001), as well as its eastern part recording the “Kibaran Event” (Tack et al., 2010) and the Irumide belt (Meinhold et al., 2011), fit well as potential sources for Mesoproterozoic and Stenian–Tonian zircon grains. These zircon-forming events occurred simultaneously with the assembly of the supercontinent Rodinia (Grenvillian orogeny). Different hypotheses have been advanced to explain this input of Mesoproterozoic and Stenian–Tonian zircon crystals (Altumi et al., 2013), including: (i) the transport of large amounts of sediment through Neoproterozoic glaciers following an original south–north transect, later reworked and deposited (Avigad et al., 2003); and (ii) a source region linked to the Transgondwanan supermountain range resulting from the East African–Antarctic Orogen and formed during the protracted Late Neoproterozoic docking of East and West Gondwana (Williams et al., 2012), involving the development of a super-fan system (Kydonakis et al., 2014; Squire et al., 2006).
Sources of Neoproterozoic detrital zircons. Neoproterozoic grains display similar age distributions in the Grandmont and Marcory formations: a main Neoproterozoic population of zircon grains (with one predominant Ediacaran group) followed by five similar peaks across the Tonian–Cryogenian interval (Fig. 4). Potential sources for these detrital zircons are located in the eastern (Saharan metacraton, Arabian-Nubian shield) and western (Trans-Saharan belt, Pan-African suture of the Anti-Atlas, early and late Cadomian arcs, and Avalonian Arc) area of the North-Gondwana margin (Fig. 5). The probability density curve of the Marcory Fm., as well as the shape and roundness of its zircon grains, implies more remote origins for them than for the Grandmont Fm. ones. Therefore, a comparison of the morphological and age differences of these zircon grains suggests an evolution of the depositional basin and it sourcing from a narrow basin being infilled by the Grandmont Fm., probably a back-arc basin resulting from the Panafrican/Cadomian orogeny, to a more evolved and opened basin infilled by the Marcory Formation, represents a more evolved and widespread basin with wider potential source areas, developed during the break-up of West Gondwana.
A composite zircon grain of the Grandmont sandstone (sample MN 1; ZR 49) demonstrates that crust with ca. 586 Ma old rocks became recycled during magmatism at ca. 567 Ma, and underlines the existence of two distinct Ediacaran magmatic events in the source area.
One should expect to identify the Ediacaran Rivernous volcanic event as reworked zircon grains in the overlying Cambrian Marcory Fm. However, this is not the case. A rapid burial of the Rivernous volcano sedimentary palaeorelief, related to high rates of sedimentation and available accommodation space due to active thermal subsidence, has been proposed for Furongian to Early Ordovician times in West Gondwana (Linnemann et al., 2011; Pereira et al., 2012) and the late Neoproterozoic–early Cambrian in the ANS (Avigad and Gvirtzman, 2009) or post-Cadomian rifting extension (Pouclet et al., in press; Von Raumer and Stampfli, 2008), which should preclude reworking of the Rivernous from distal (northern) to proximal areas (southern Montagne Noire).
6 Conclusions
The Ediacaran–Cambrian boundary has been confidently identified within error, based on U–Pb zircon dating, into the Rivernous Fm. of the northern Montagne Noire. The Rivernous volcanic event is indicated to be the lateral equivalent of the Sériès Tuff of the Axial Zone. This fits well with a latest Ediacaran depositional age (ca. 574 Ma) estimated with detrital zircon U–Pb geochronology for the underlying Grandmont Fm. The latter should be considered as a time-stratigraphic equivalent of the acritarch-bearing “Schist X” Fm. of the Axial Zone (Supplementary data, Fig. S3).
U–Pb analysis of the detrital zircons from the Ediacaran Grandmont Fm. and the Cambrian Series 2 Marcory Fm. suggests a change over time in the sourcing. The Ediacaran sediments were deposited in a narrow back-arc basin far from the influence of far cratonic sources, whereas Cambrian Series 2 detrital sediments were derived from mature source rocks involving the denudation of different Gondwanan cratons.
Acknowledgements
The authors thank constructive criticism made by O. Blein and B. Laumonier, and founding from the RGF program of the French Geological Survey (BRGM). This paper is a contribution to project CGL2013-48877-P from Spanish MINECO.