The Ordovician-Silurian boundary (late Katian ...

The Ordovician-Silurian boundary (late Katian-Hirnantian) of western Anticosti Island: revised stratigraphy and benthic megafaunal correlations Paul Copper1, Jisuo Jin2, and André Desrochers3 2

1 Loupicoubas, 46220 Prayssac, France Department of Earth Sciences, Western University, London, ON, N6B 56A, Canada 3 Department of Earth Sciences, University of Ottawa, K1N 6N5, Canada

email: [email protected]

ABSTRACT: The type sections of the Ellis Bay Formation are revised to incorporate recent stratigraphic correlations east to west on Anticosti Island. This is one of the most complete tropical carbonate sequences spanning one of the major Phanerozoic mass extinction episodes. A richly fossiliferous benthic fauna faithfully records the change-over from the Ordovician (Richmondian, late Katian), the mass extinction around the boundary, and the recovery within the earliest Silurian (Rhuddanian, Llandovery). Critical in this revision is the Hirnantian shelly fauna that begins at or close to the base of the Ellis Bay Formation at the west end (as revised herein), and ends at the top of the reefal Laframboise Member, the O/S boundary as used herein. Equivalent strata at the east end contain Hirnantia sp. at the base of the Prinsta Member, as the basal Ellis Bay Formation, and Hirnantia sagittifera in reef-capping beds at the top of the Laframboise Member in the central area of the island. The Ellis Bay Formation is rich in typical Hirnantian brachiopods such as Eospirigerina and Hindella. The earliest Silurian recovery brachiopod fauna is usually small-shelled, composed of both Ordovician hold-over taxa (ca 30%) and new arrivals such as the pentameride Viridita, athyridide Koigia, and atrypide Zygospiraella. In this study, three new members are proposed for the 80–90m thick Ellis Bay Formation at the west end of Anticosti Island, beginning with the basal shales and limestones of the Fraise Member (overlying a recessive shaly unit of the Katian Vaureal Formation), followed by the Juncliff Member resistant limestones, and the overlying limestones and shales of the Parastro Member. The uppermost two members are correlated directly with the Lousy Cove and Laframboise members at the east end, where the same sequence is thinner and contains several discontinuities. The top of the reef-capping grainstone beds above the reefal Laframboise Member marks the major end-Ordovician extinction of Hirnantian brachiopods, stromatoporoids, corals, crinoids and nautiloids, signaled by the return to background δ18O and δ13C values. The overlying Rhuddanian Becscie Formation comprises a lower Fox Point Member of thin, evenly bedded limestones, with a low-diversity but high-abundance Rhuddanian brachiopod fauna, and an upper Chabot Member of irregularly bedded coralline, non-reefal limestones, marking the appearance of typical Silurian benthic faunas.

INTRODUCTION

One of the five most significant Phanerozoic Mass Extinction Events (MEEs) was close to the end of the Ordovician Period. Despite increasing interest in the end Ordovician MEE, there is considerable controversy about how to interpret the global picture, and its recovery therefrom, especially within the final stage of the Ordovician, the Hirnantian. Significant controversy exists also on a regional scale, such as the biostratigraphic constraints on the faunal turnovers in eastern North America. The key carbonate sequence of Anticosti Island (text-fig. 1), largely uninterrupted across the O/S boundary, was located in a southern tropical (~25° paleolatitude) carbonate ramp to shelf setting at the west end of Anticosti Island. There, O/S boundary strata are exposed continually on the tidal flats and in the adjacent shoreline bluffs. The complete western section stretches from Anse aux Fraises to Pointe Laframboise for some 4km, and also at both sides of Ellis Bay, thus intersecting the upper Vaureal, Ellis Bay and Becscie formations. Three different scenarios have been proposed for the O/S extinctions and boundary on Anticosti: (1) a terminal extinction within the Hirnantian, marked by isotopic anomalies, followed by a pre-Silurian recovery with standard marine isotopic signatures (e.g., Brenchley 1982); (2) a final extinction within a reefal episode in the earliest Silurian (Barnes 1981; partim Petryk 1981); and

Stratigraphy, vol. 10, no. 4, pp. 213–227, text figures 1–6, 2013

(3) a final extinction of the classic Richmondian (late Katian) fauna at the top of a reefal episode (Schuchert and Twenhofel 1910), followed by an earliest Silurian dwarfed and low diversity recovery fauna in the Becscie Formation (Copper 2001). The first interpretation based on 8–9m of Anticosti Hirnantian section (Brenchley 1982), with the rest of Katian age, has been the prevailing view in the literature. For the Anticosti succession, there is still much controversy about the significance of δ13C and δ18O excursions across the O/S boundary, the application of sequence stratigraphy to the Anticosti sections, and the detailed correlation between the microfaunas (conodonts, chitinozoans, and acritarchs) and the richly fossiliferous shelly and coralline megafaunas of the carbonate platform. The graptolite fauna is indeterminate to date, as graptolites are scarce and fragmentary in the shallow marine carbonate settings, as they are in the condensed Estonian Hirnantian section. There are no black shale or other organic-rich strata in the upper Katian through middle Rhuddanian on Anticosti Island. THE ORDOVICIAN-SILURIAN BOUNDARY CONUNDRUM

The Late Ordovician (late Katian) through Early Silurian (Rhuddanian) strata crossing the O/S boundary record one of the five major Phanerozoic MEEs. In much of the marine setting, there was a major and widespread hiatus at this horizon, and thus uninterrupted sections that log a complete picture are 213

Paul Copper et al.: The Ordovician-Silurian boundary (late Katian-Hirnantian) of western Anticosti Island

relatively rare. These events were initially ranked second in severity after the end-Permian extinctions by Sepkoski (1982), and then downgraded to fourth in severity more recently by Alroy (2008, 2010a, 2010b). At the time of the mass extinctions towards the end of the Ordovician, the expanded south Polar ice cap and surrounding ice shelves were situated in Gondwanan north Africa (Ghienne et al. 2007; Scotese 2010; Ghienne 2011). This drove Earth towards an intense icehouse setting lasting ca 2–3 myr, interrupted by interglacial warmings, during the late Katian and mid-Hirnantian. Their duration was roughly equivalent to the Pleistocene icehouse. Ghienne et al. (2007) recognized two first-order cycles within the Hirnantian alone, separated by a major mid-Hirnantian deglaciation (or interglacial episode, marked by sealevel highstand). Thus there were at least three latest Katian through end Hirnantian glacial events, accompanied by sealevel drawdowns. The concept of a glacially driven global mass extinction at or near the end of the Ordovician was first proposed by Sheehan (1973), and expanded by Brenchley (1988). Other authors have intimated only a single terminal Ordovician extinction (Orth et al. 1986; Brenchley 1988, 2004; Finnegan et al. 2011). On Anticosti there were at least two major extinctions, one at or close to the top of the Vaureal Formation (possibly coinciding with the top of the last Vaureal Paleofavosites-Palaeophyllum reefs just below the Schmitt Creek Member (Long and Copper 1987), and the other at or near the top of the Laframboise Member of the Ellis Bay Formation (Copper 2001). There was a short reefal episode within the lower Ellis Bay Formation in the Vaureal River area that intimates a possible lowstand: the faunas in this reefal episode are different from those of the Katian and late Hirnantian (Lake 1981; Long and Copper 1994). Glaciation was likely the main driver for the multiple episodic extinctions, as suggested by δ18O isotopic signatures (Laporte et al. 2009). There is no evidence for an end Ordovician extraterrestrial impact based on the rock record of Anticosti Island (Orth et al. 1986) or elsewhere. Concomitant with cooling at this time, the southern margin of the Laurentia plate, along which Anticosti was a shallow carbonate platform to ramp, was located at about 25ºS latitudes (Nestor et al. 2010; Cocks and Torsvik 2011; Jin et al. 2013). This generated easterly currents, influenced by seasonal typhoons, not unlike much of the Great Barrier Reef today and in similar subequatorial latitudes. Anticosti displayed reef development in the late Katian, a reefal episode within the early and also final Hirnantian (Copper 1989; Copper 2001), and a late Aeronian coral patch reef expansion (Copper and Jin 2012). Laurentia was separated from the fused Baltica-Avalonia plates by the relatively narrow and elongated Iapetus ocean, some 1000km wide to the south (Cocks and Torsvik 2011). The IUGS global O/S boundary working group summarized the controversies after field meetings at the global standard stratotype, in the graptolitic shales of Dob’s Linn in Scotland (reviewed by Williams 1988), with worldwide boundary data compiled by Cocks and Rickards (1988). These followed the Anticosti Boundary Working Group meeting on Anticosti (Lespérance 1981), where an alternate stratotype in shallow water carbonate sections was examined. The Dob’s Linn section (at ~30ºS paleolatitudes; see Zalasiewicz and Tunnicliff 1994; Underwood et al. 1994; Verniers and Vandenbroucke 2006) exposes the late Katian through Rhuddanian Hartfell and Birkhill black shales spanning the pacificus, extraordinarius, persculptus and acuminatus graptolite biozones. The section is condensed (4–6m thick) and tectonically overturned and faulted. On the other hand, the Anticosti section is undisturbed, essentially horizontal in its original dip of 1-2º, and was 214

deposited at the southern tropical equatorial belt in the Late Ordovician and earliest Silurian (Scotese 2010). Moreover, it has a rich invertebrate megafauna, and diverse microfossils such as conodonts (Nowlan 1982; Barnes 1988), ostracodes (Copeland 1973), chitinozoans (Soufiane and Achab 2000; Achab et al. 2012), and acritarchs (Vecoli et al. 2011; Delabroye et al. 2012). It is on the 80–90m thick Anticosti section at the west end of the island that we focus this paper. The O/S boundary outcrops extend for some 200km along strike from west to east, on the island, including numerous tidal flat, coastal bluff, river and, more recently since the 1990s, new road outcrops. More than 100 GPS-defined localities for the Ellis Bay Formation are used for this study, supplemented by Katian-Hirnantian boundary sections that yield key information on multiple faunal turnovers from the Late Ordovician through Early Silurian. BACKGROUND

Anticosti Island was first geologically mapped and measured in 1856 by James Richardson, the pioneer Paleozoic geologist for the then recently founded Geological Survey of Canada, established in 1845. Richardson (1857), in the ensuing publication, assigned the Paleozoic strata on Anticosti to six Divisions, from A to F (with A to C later assigned to the ‘Lower Silurian’, now the Ordovician), and the rest ‘Upper Silurian’ (now Llandovery). Logan (1863, p. 298) placed the strata overlying his ‘Hudson River formation’ (later the Vaureal Fm), as exposed at the west end of the island, in his basal ‘Anticosti group’. This included both what were later to become ‘Gamachian’, as well as the Lower Silurian beds of the Becscie Formation. Logan (1863, p. 298-300) divided the equivalent of Richardson’s ‘Division C’ (what was later to become the Ellis Bay Formation) into 14 units, of which unit 11 was Richardson’s coral bed at Pointe Laframboise, and, unit 12, a ‘grey limestone with argillaceous partings 62 feet thick’ (ca 19m = Becscie Formation). Historically, it is thus just over a century ago that Schuchert and Twenhofel (1910) laid down the formation names for one of the finest O/S boundary sections worldwide. At that time, Schuchert and Twenhofel also coined the stage (and series) name ‘Gamachian’ for post-Richmondian, latest Ordovician, and pre-Silurian strata, based on the Anticosti succession faunas. Lapworth (1879), who founded the Ordovician System, set the standard graptolite zones at Dob’s Linn Scotland (see Underwood et al. 1994 for chemostratigraphy, and Verniers and Vandenbroucke 2006 for chitinozoans), as key zones for deep water successions. With the scarcity of graptolites in the shallow water carbonates and shales of Anticosti (Riva 1969, 1988; Melchin 2008), the Lapworth sequence has remained difficult to correlate with both the Anticosti and Estonia successions. Biostratigraphy of the western sections of Anticosti Island can be traced back to Richardson (1857), who collected fossils at prominent outcrops around Ellis Bay. The fauna was examined and described by Billings (1857, 1862, 1865), who named a number of new fossil genera that were at the time unknown elsewhere in North America. Billings sent a number of Anticosti specimens to various museums around the world, including the Natural History Museum (London), the Musée d’Histoire Naturelle (Paris), the Humboldt Museum (Berlin), and even the La Plata Natural History Museum (Argentina). This quickly attracted others such as Shaler and Hyatt from the Museum of Comparative Zoology at Harvard (1865), and Schuchert and Twenhofel from Yale University (1910), the latter authors naming new formations that are now established in the literature. Twenhofel garnered further materials in 1914, completed more fieldwork in 1919, and subsequently published his massive volume, the ‘Geology of Anticosti Island’ (1928), describing

Stratigraphy, vol. 10, no. 4, 2013

TEXT-FIGURE 1 Geologic map of Anticosti Island, and satellite imagery of the critical west end stratigraphy from Anse aux Fraises (north) to Pointe Laframboise (south). contrasting the resistant and recessive weathering strata of the Ellis Bay Formation. The west side of Ellis Bay itself is at the right side of the satellite image. The megafaunally defined O/S boundary is at the top of reef-capping grainstone beds above the Laframboise reef and inter-reef strata in the lower figure, and the Katian-Hirnantian boundary is located in the Vaureal Formation-Fraise Member transition interval. The soft weathering, shaly uppermost Vaureal Formation of the west end defines the Anse aux Fraises recessive notch on the tidal flats.

215


TEXT-FIGURE 2 Approximate correlation of the west and east end stratigraphy of Anticosti Island. West end measured stratigraphic section based on outcrops from Anse aux Fraises to Pointe Laframboise: the section is roughly twice as thick as that of the northeast coast. The Fraise, Juncliff and Parastro members are roughly equivalent to the Prinsta and lower Lousy Cove members of the northeast coast, but much thicker in the west. Correlation of the Lousy Cove Member with the type northeast coast section is based on lithology and benthic faunas. Diagram modified from Copper (2001).

more new taxa and attempting wider biostratigraphic zones and correlations. THE ELLIS BAY FORMATION

Despite its virtually pristine, undeformed, and stratigraphically complete sequence, the boundary sections of the Ellis Bay Formation have been controversial (text-fig.2). In part this has been so because of the island’s remote location and poor accessibility, so that few have visited, or seen the geology only briefly. This has meant that there have been several interpretations for placement of the critical O/S boundary, or how thick the Hirnantian component of the section has been. 216

Richardson (1857) measured his Division C, effectively what was later to become the Ellis Bay Formation, as ‘167 feet, 6 inches’ thick (ca 51m). The base of ‘Division C’ of Richardson (1857, p. 215) was described as 3.8m of recessive shales and limestones at the base of Junction Cliff. The top of Division C was marked by ‘the last bed occurs at Pointe Laframboise, a coral limestone measured at 4 ft 6 in thick’ (1.4m; Richardson 1857, p. 216). Richardson also found this same western reefal unit near Cap Henri and close to Cap à l’Aigle on both sides of Ellis Bay: this is the Laframboise Mbr of modern usage. Though Richardson did not clearly identify the base of Division C around Anse aux Fraises, where he did not examine it as exposed at low tide, the top of the reefs marks the end of the Ordovician sequence. At the


TEXT-FIGURE 3 Range chart of critical benthic brachiopod species across the Ordovician-Silurian boundary, sorted by orders. The maximal thickness of the Ellis Bay Formation at the west end is 80-90 m: the Fox Point member here is ca 35m thick, and ends ca 300m NW of Pointe à l’Ours. Note that much of the late Katian (Richmondian) fauna disappears at the top of the Vaureal Formation, with some hold-over taxa into the Ellis Bay Formation. Within the Ellis Bay Formation there is an early Hirnantian cluster of taxa (including Hirnantia sp. at the base), a mid-Hirnantian suite, and a late Hirnantian group, including Hirnantia saggitifera, in beds capping and flanking the Laframboise reefs. Within the Hirnantian there was relatively rapid evolution of Hirnantian species, especially in the newly arrived spire-bearers such as atrypides and athyridides. Some of the ecologically selective, non-reefal brachiopod species disappear below the final Laframboise Member. The basal few meters of the Fox Point Member sees the loss of further Katian genera and species: this is also marked by relative dwarfism, low diversity and introduction of Silurian pioneering genera such as Zygospiraella, Viridita and Koigia that mark the recovery after the O/S Mass Extinction.

northeast end of the island, Richardson correlated the sandstones at Cape James and all sections as far as ‘the old provision post at Fox River’ in Division C, which would today encompass the upper Vaureal through lower Becscie formations, thus spanning a far greater unit of time (Richardson 1857, p. 218). However, Richardson (1857, p. 217) also noted that on the northeast coast. ‘in the bight of Prinsta Bay...they [the sediments of Division C] ... succeed the sandstones’ [of his Division B and the cliffs at Cape James], ... they crown Table Head and come to the level of the water on the east side of it’. The last location, ‘the lower level of the water’, is Lousy Cove, a small embayment east of Table Head lighthouse erected in 1915, used to ferry supplies to the light house at the top of Table Head cliff. This locality, and the sandstone/shale contact at Prinsta Bay, marked the base of the Prinsta Member of Long and Copper (1987) and was featured in Cameron and Copper (1994) and Copper et al. (2011). Billings (1857, p. 252), who analysed the fossils collected by Richardson, remarked that Richardson’s ‘Division C’ contained only about one third of the faunas of Divisions A and B (the Vaureal Fm) that occur in Division C (the Ellis Bay Fm), and ‘the others do not appear any more and probably become extinct’. Billings was thus the first to notice that the top of Division B marked extinction of two thirds of the Richmondian (or ‘Hudson River’), i.e. upper Katian fauna. Billings (1857, p. 253) also noticed that the fauna of the lower part of the succeeding Division D (the Becscie Fm) was impoverished in species, thus indirectly confirming a second extinction for the end Ordovician. Schuchert and Twenhofel (1910) coined the name ‘Ellis Bay Formation’, considering it as the final ‘Gamachian Stage’ of the Ordovician, younger than the classic Richmondian of North America. They partitioned the formation into 11 faunal zones, the base being unit 1 of Richardson’s Division C ‘met with at Junction Cliff’ (Schuchert and Twenhofel 1910, p. 701) and with

a total thickness of 180 ft (55m). At the top of their Ellis Bay Formation was their ‘first marked coral reefs’ assigned to zones 10 and 11 at the close of the Ordovician (Schuchert and Twenhofel 1910, p. 685, 704). Their cumulative thickness for zones 10 and 11 was 20 ft (6.1m), which is approximately the thickness of the reefs through the Pointe Laframboise outcrops at the west end. They did not describe the nature of these beds, so it is likely that zones 10 and 11 included both reef, reef-capping, and interreef strata. Schuchert and Twenhofel (1910, p. 505) specifically mentioned that ‘the line between the Cincinnatic and Siluric periods or systems [is] between Richardson’s and Logan’s zones C 11 and C12’. Thus the top of their Ordovician was the top of the coral reefs, matching the evaluation of Richardson (1857), and subsequently the top of the reefal Laframboise Member of Long and Copper (1987). Confusingly, Twenhofel (1928, p. 48) re-described and renumbered the units of the Ellis Bay Formation and called his new ‘Zone 9’ the ‘coral limestone’, drawing the top of the formation some 30 ft (9m) higher above the reefs, in his Zone 10. Bolton (1961) identified ‘50 to 120 feet of strata’ (15–37m) above the ‘biohermal zone’ of Twenhofel (1928) as belonging to the Ellis Bay Formation. In 1972 Bolton divided the formation into just six members, with his Member 6 having the reefs at the base, overlain by another 15–37m of ‘semi-lithographic limestone’ with ripple marks, thus following Twenhofel (1928), and not Schuchert and Twenhofel (1910). The ripple marked units with nautiloids were identified as Ordovician by Dixon (1970), but Holland and Copper (2008) assigned these nautiloids to the Lower Silurian Becscie Formation. The top of this correlation of the ‘Ellis Bay Formation’, somewhere between 9–40m above the reefs, is impossible to define in the field either lithically, or from aerial photos using marker beds or units. It is also impossible to define macrofaunally: the final extinctions occur at, or near, 217


TEXT-FIGURE 4 Outcrop views of the Fraise and Juncliff members, Ellis Bay Formation, at the west end of Anticosti. (a) Aerial view looking northwest with Anse aux Fraises in the distance, and Junction Cliff (foreground right): red asterisks mark the lower and upper contacts of the Fraise Member over a distance of ca 1km; (b) air photo of the west end tidal flats showing the lower and upper contacts of the Fraise Member marked by red asterisks (scale bar = 400): much of this new member has been assigned to the Vaureal Formation in the past (Petryk 1981a). Note the change to resistant limestone lithology in the middle of the member; (c-d) views of Junction Cliff looking south, exposing ca 3m of the recessive shaly Fraise Mbr at the base, and ca 18m of the overlying resistant Juncliff Mbr micrites (contact with yellow asterisk); (e) view of the tidal flat lithology at Anse aux Fraises, defining the lower contact of the Fraise Member with the recessive underlying Vaureal Formation; (f) tidal flat outcrop at the south end of Anse aux Fraises showing the lower bedding plane surfaces of the Fraise Member marking the FAD of the conodont Gamachignathus and the Hirnantian fauna.

218


TEXT-FIGURE 5 Air photo, at high tide, of the ca 2km long tidal flat section at the west end of Anticosti, from Parastrophinella Bluff to Pointe Laframboise, illustrating the contact of the recessive Parastro Member with the underlying Juncliff Member and the overlying Lousy Cove and Laframboise members. Parastrophinella Bluff is at the north end, the type locality of Parastrophinella reversa; (b-c) the basal resistant limestones of the Lousy Cove Mbr looking west (see red asterisk in 5a) are marked by a gentle tilting of strata; (d) bluff at the west side of Ellis Bay, exposing the Parastro Member near its base (locality A1171); (e) bluff at the west side of Ellis Bay, towards Cap Henri, exposing a good section of the resistant weathering Lousy Cove Member, with the lower contact at the base of the bluff (locality A1174).

the top of the reefal Laframboise Member, with a new, generally dwarfed, and low-diversity Silurian fauna in the more resistant, rhythmic Becscie limestones directly above. Thus the basal Becscie, as herein interpreted, defines the faunal recovery after the multiple O/S mass extinction episodes that took place from the late Katian through the end Hirnantian. To solve the boundary problems between the Ellis Bay and Becscie formations, and the O/S boundary, a redefined demarcation was

adopted by Petryk (1981b, p. 19) at the top of the ‘biohermal zone’... ‘recognizable and mappable over the whole extent of Anticosti’. This reverted the boundary to the top of Division C of Richardson (1857), and the definition of Schuchert and Twenhofel (1910), back to its original limits. This was illustrated in a diagram by Petryk (1981a, fig. 8 and 1981b, fig. 10). However, in this same figure, the reefs are shown at the top of Petryk’s zone 7 of the Ellis Bay Formation, and the inter-reef beds and fossiliferous reef cap were placed in the lower 1.5–2m of the Becscie Formation. In 219


these figures Petryk (1981a, 1981b) also stated that ‘Twenhofel’s 1928 Ordovician-Silurian boundary’ was ‘9.5 to 10 m’ above the reefs. Since the reefs, the inter-reef and the immediate reef-capping coral, stromatoporoid, crinoid and brachiopodrich beds contain the same faunas as the reefs themselves, it is difficult to understand how they can be assigned to two different formations and two different systems, one in the Ordovician and the other Silurian. Moreover, the immediate fossiliferous reefcapping and inter-reef beds of the Laframboise Member in the Natiscotek outcrops contain well-preserved, calcitic shells of characteristic Hirnantia sagittifera (see Zhan and Jin 2008). The top of the Ellis Bay Formation, defined by the megafauna, is equivalent to the top of the Laframboise reef and encriniterich and shelly inter-reef and reef-capping strata, as originally stated by Schuchert and Twenhofel (1910). This horizon also defines the O/S boundary in terms of the end of the Gamachian (or Hirnantian) fauna and the beginning of the Silurian fauna. Such an upper boundary (see text-fig. 3) is clearly evident in the brachiopods (Jin and Copper 1997, Jin and Zhan 2008, Copper 2001), chitinozoans (Achab et al. 2012), and conodonts (McCracken and Nowlan 1988). At present, it also matches the sparse graptolite data (Melchin 2008; Melchin et al. 2003). The base of the Prinsta Member at the northeast coast, and the base of the Ellis Bay Formation there, as modified by Desrochers et al. (2010), exposed at Lousy Cove, contains another species of Hirnantia (Jin and Zhan 2008). This suggests that both the base and top of the Ellis Bay Formation on Anticosti contain the index genus Hirnantia, although this older species has yet to be found on the tidal flats of the Ellis Bay Formation at Anse aux Fraises in the west. Desrochers et al. (2008) and Delabroye et al. (2011) confirmed that, in the east, the Ellis Bay Formation begins with the base of the Prinsta Member at Prinsta Bay, and that the formation is thinner and contains several hiatal surfaces. Significantly, Hirnantia sp. at the base of the Prinsta Member confirms that the Velleda and Grindstone members of Long and Copper (1997) are best allocated to the underlying Vaureal Formation. Finnegan et al. (2011), in contrast, considered that only the uppermost few meters of the Ellis Bay Formation at the northeast end could be considered as Hirnantian, following Brenchley et al. (2003). Using conodonts, Barnes and McCracken (1981, Table 1) and Barnes (1988, Fig.3) respectively placed their reefal member 6 or member 7 within the upper Ellis Bay and the lower Becscie formations, thus with the O/S boundary somewhere within the reefs. Barnes (1981, p. 215) iterated that ‘the major extinction event in conodonts occurs within latest Ordovician time and not at the new systemic boundary’. Thus effectively the conodont biostratigraphy follows lithic correlation errors first established by Twenhofel (1928) and continued by Bolton (1961, 1972). There would be no such problem with any faunas, including the conodont data, if the O/S boundary, and the division between the Ellis Bay and Becscie formations were simply settled at the top of the reef-capping beds above the Laframboise reefal and inter-reefal strata. That was a strategy used in a wide range of papers using macrofaunas on Anticosti (e.g. Cocks and Copper 1981; Copper 1981; Johnson et al. 1981; Dixon et al. 1986; Long and Copper 1987; Copper 1989, 2001; Jin and Copper 2008; Jin and Zhan 2008- and text-fig. 3 herein). The Ellis Bay (Laframboise Member) reefs and inter-reef strata also contain a Hirnantian stromatoporoid fauna similar to that of the Pirgu and Porkuni formations of Estonia (Nestor et al. 2010), the last of the aulaceratid stromatoporoids (Copper et al. 2011), and the final Ordovician tabulate corals (Dixon 1974, 1986, Dixon et al. 1986). The rugose coral fauna crossing the O/S boundary (McLean and Copper, 2013), corroborates the stromatoporoid 220

and tabulate coral data. Colonial rugosans are large and common in the Laframboise (Hirnantian) reefs, but in the succeeding Rhuddanian Fox Point Member consist only of solitary forms. REVISED WESTERN STRATIGRAPHY OF THE ELLIS BAY FORMATION

The Ellis Bay Formation on the west coast is herein newly subdivided for the western facies on Anticosti Island, in the area between Anse aux Fraises (westernmost outcrops) and the Laloutre River area about 60km to the east (text-figs. 1, 2). It has been difficult to correlate precisely with the far eastern O/S contact outcrops about 190km from Anse aux Fraises, where thick sandy units were arbitrarily assigned by Schuchert and Twenhofel (1910) to their Ellis Bay Formation (text-fig. 3). These sandy units have a sparse megafauna dominated by molluscs, lacking age-diagnostic brachiopods, stromatoporoids and corals. The 1910 scheme was followed by Long and Copper (1987), who maintained the assignment of the two basal sandy units to the Ellis Bay Formation, which units they named the Grindstone and Velleda members. The overall thickness of the basal sandy members plus the eastern shaly and limestone units above matched the thickness of the Ellis Bay Formation as named from the west end (ca 80–90m thick). The base of the Ellis Bay Formation at the east end was then correlated with the top of the shaly Schmitt Creek Member of the Vaureal Formation, a lithic correlation extended to the west end, where a similar shaly unit of the Vaureal Formation underlies limestones and shales of the type Ellis Bay Formation. This correlation was followed in a geologic map and stratigraphic column revision for the west end by Jin and Copper (1997), that is now shown to be incorrect, and is herein revised. SEQUENCE STRATIGRAPHY

The east-west trending, ~200km long, outcrop belt of the Ellis Bay Formation on Anticosti Island was recently re-examined using a sequence stratigraphic approach (Desrochers et al. 2010). This outcrop belt is slightly oblique to the paleoshoreline with nearshore siliciclastic-dominated facies restricted to the eastern end of the island and more offshore carbonate-dominated facies present along the central part and western ends. Five transgressiveregressive (TR-1 to TR-5) sequences were recognized allowing a more precise intraformational correlation of individual Ellis Bay members or units. In the easternmost sections, relatively thin T-Rsequences are locally bounded by disconformities. In the central and western sections, the thicker T-Rsequences have more symmetrical, transgressive-regressive components reflecting a greater subsidence compensated by a higher carbonate sediment supply. The tops of TR-4 and TR-5 sequences (sensu Desrochers et al. 2010) are, however, truncated, from east to west, by regional erosional surfaces corresponding to the base and the top of the distinctive Laframboise Member. At Pointe Laframboise or along the Salmon River, the TR-5 sequence or Laframboise Member is sharply overlain, or draped over, by a thin transgressive bioclastic lag (also called the capping reef facies of the Laframboise Member by other workers) passing rapidly upward into thin-bedded, distal ramp tempestites at the base of the Becscie Formation. Based on integrated studies of chitinozoans and acritarchs, sequence stratigraphy and carbon chemostratigraphy (Delabroye et al. 2011; Desrochers et al. 2010; Achab et al. 2011, 2012), all five T-R sequences defined in the Ellis Bay Formation are likely Hirnantian in age. In terms of sequence stratigraphy, chitinozoan biostratigraphy and chemostratigraphy, the O/S boundary may be placed within a transgressive phase, a few meters above the base of the grainstone beds capping the Laframboise reefal and interreefal strata. The terminal Ordovician megafaunal extinction on


TEXT-FIGURE 6 The Laframboise Member type outcrops on the tidal flats at the west end and in Ellis Bay. (a) southward aerial overview of the tidal flats at low tide, with wave breakers marking the outer edge of the flats (ca 1km wide) and the chief contact markers for the formations and members at the megafaunally defined O/S boundary; (b) the tidal flat at the type Pointe Laframboise section: the oncolite or basal unit of the Laframboise is at the north edge (asterisk) and the reef exposures extend for about 1300m [ca 8m thickness: scale bar = 150m; zoom in to see patch reefs]; (c) tidal flats at the west side of Ellis Bay with the Laframboise Member, and megafaunally defined O/S boundary (asterisks); note shaly interval, thin oncolite bed at the north end, and recessive reef capping beds below the Becscie Formation (scale bar = 150m); (d) tidal flats at the east side of Ellis Bay intersected by a minor normal displacement - several grainstone and patch reef units follow each other at this side and the bedding is irregular (scale bar = 150m); note that in each tidal flat air photo, the Fox Point Member consists of very thin, evenly bedded, rhythmic, hard grainstones that form a distinctive pattern on the tidal flats; (e, f) side view of a small patch reef of the Laframboise Member at the west side Pointe Laframboise, showing the top of reef-capping grainstone beds (red asterisk), about 1m here above the Laframboise reef and inter-reefal strata (yellow dotted line). The Ellis Bay-Becscie formational contact varies from section to section along the 200km outcrop belt on Anticosti Island and is subject to interpretation. At Pointe Laframboise, this lithostratigraphic contact can be placed at the red asterisk (sensu Copper and Jin) or the yellow dotted line (sensu Desrochers).

221


Anticosti occurs also at the top of these grainstone beds (see section on Laframboise Member and text-fig. 6e and f). Fraise Member (new)

The Fraise Member, ca 20m thick, includes interbedded calcarenites and thin shales towards the base, with increasing dominance of recessive weathering shales to the top. It is named after Anse aux Fraises at the west end, where the unit crops out on the tidal flats towards Junction Cliff (text-figs. 4a,b). The upper and lower contacts of the new member are readily identified on air photos (see text-fig.1). The base of the unit is at the base of the first resistant limestone unit exposed on the tidal flats at Anse aux Fraises (text-fig. 4b), as marked in Jin and Copper (1997), overlying a still unnamed recessive unit of the Vaureal Formation, that is perhaps equivalent to the Prinsta Member of the northeast coast. This poorly exposed shaly unit of the Vaureal Formation was initially correlated with the Schmitt Creek Member on the northeast coast (Long and Copper 1987). This has turned out to be incorrect as the Schmitt Creek Member lies well below the Vaureal-Ellis Bay contact, and below the sandy Grindstone Member. The top of the Fraise Member is exposed at Junction Cliff (named by Richardson in 1856 = Cap de la Vache qui Pisse of Petryk 1981a), and marks a clear break between soft grey shales and the thin- to medium-bedded, resistant micritic limestones of the Juncliff Member above (text-fig. 4c). However, Petryk (1981a,b) assigned only