Terminal Ordovician Carbon Isotope Stratigraphy and Glacioeustatic Sea-Level Change Across Anticosti Island (Québec, Canada)

Home , Belonechitina, Hirnantian, Katian

Terminal Ordovician carbon isotope stratigraphy and glacioeustatic sea-level change across Anticosti Island (Québec, Canada)

David S. Jones1,†, David A. Fike1, Seth Finnegan2, Woodward W. Fischer2, Daniel P. Schrag3, and Dwight McCay1 1Department of Earth and Planetary Sciences, Washington University in St. Louis, 1 Brookings Drive, St. Louis, Missouri 63130, USA 2Division of Geological and Planetary Sciences, California Institute of Technology, 1200 East California Boulevard, MC 100-23, Pasadena, California 91125, USA 3Department of Earth and Planetary Sciences, Harvard University, 20 Oxford Street, Cambridge, Massachusetts 02138, USA

ABSTRACT approach provides evidence for the diachro- picture (Runkel et al., 2010). The biological crisis neity of the oncolite bed and Becscie lime- that accompanied the glaciation was the second Globally documented carbon isotope ex- stones; the former transgresses from west greatest mass extinction in the marine fossil cursions provide time-varying signals that to east, and the latter progrades from east record (Sheehan, 2001; Bambach et al., 2004), can be used for high-resolution stratigraphic to west. The sea-level curve consistent with depleting the enormous diversity of marine or- correlation. We report detailed inorganic and our sequence-stratigraphic model indicates ganisms that had evolved during the Ordovician organic carbon isotope curves from carbonate that glacioeustatic sea-level changes and the Radiation (Webby et al., 2004). Marine carbon- rocks of the Ellis Bay and Becscie Formations positive carbon isotope excursion were not ate rocks deposited at the time preserve records spanning the Ordovician-Silurian boundary perfectly coupled. Although the start of the of a global disturbance to the carbon cycle rep- 13 on Anticosti Island, Québec, Canada. Strata of isotope excursion and the initial sea-level resented by positive excursions in I Ccarb and 13 the Anticosti Basin record the development of drawdown were coincident, the peak of the I Corg (Marshall and Middleton, 1990; Brench- a storm-dominated tropical carbonate ramp. isotope excursion did not occur until after sea ley et al., 1994; Underwood et al., 1997; Wang These strata host the well-known Hirnantian level had begun to rise. Carbon isotope values et al., 1997; Ripperdan et al., 1998; Kump et al., positive carbon isotope excursion, which at- did not return to baseline until well after the 1999). Some of the sedimentary records also host tains maximum values of ~4.5‰ in carbonate Anticosti ramp was reÁ ooded. The sea-level– geochemical evidence of global cooling and/or 13 carbon of the Laframboise Member or the I Ccarb relationship proposed here is consis- ice-sheet growth represented by parallel positive 18 Fox Point Member of the Becscie Formation. tent with the “weathering” hypothesis for the excursions in I Ocarb (Marshall and Middleton, 13 The excursion also occurs in organic carbon, origin of the Hirnantian I Ccarb excursion. 1990; Long, 1993; Brenchley et al., 2003). 13 13 and I Ccarb and I Corg values covary such Although much progress has been made in 13 13 13 that no reproducible ) C (= I Ccarb – I Corg) INTRODUCTION end-Ordovician correlation and carbon isotope excursion is observed. The most complete chemostratigraphy (Finney et al., 1999; Kump stratigraphic section, at Laframboise Point One of the largest mass extinctions of the et al., 1999; Brenchley et al., 2003; Melchin in the west, shows the characteristic shape of Phanerozoic took place at the close of the Ordo- and Holmden, 2006; Kaljo et al., 2008; LaPorte the Hirnantian Stage excursion at the global vi cian Period (Sepkoski, 1981; Brenchley et al., et al., 2009; Yan et al., 2009; Zhang et al., 2009; stratotype section and point (GSSP) for the 2001; Sheehan, 2001; Bambach et al., 2004), Ainsaar et al., 2010; Desrochers et al., 2010; Hirnantian Stage in China and the Silurian an interval also marked by the development of Young et al., 2010), uncertainty remains about System in Scotland. We therefore suggest extensive continental ice sheets on the south the causal relationship among glaciation, mass that the entire Hirnantian Stage on Anticosti polar super continent Gondwana (Hambrey extinction, and perturbation to the carbon cycle. Island is conÀ ned to the Laframboise and and Harland, 1981; Hambrey, 1985; Ghienne , Delabroye and Vecoli (2010) recently reviewed lower Fox Point Members. 2003; Le Heron et al., 2007; Le Heron and some of the outstanding issues in Hirnantian By documenting discontinuities in the Dowdeswell, 2009; Le Heron and Howard, 2010; event stratigraphy, emphasizing shortcomings archi tecture of the carbon isotope curve at Loi et al., 2010). This glaciation and the possible in the current biostratigraphic and chemostrati- multiple stratigraphic sections spanning a Early Silurian glaciation (Caputo and Crowell, graphic correlations. Although many pieces of proximal to distal transect across the sedi- 1985; Grahn and Caputo, 1992; Díaz-Martínez evidence suggest that the glaciation, extinction, mentary basin, we are able to reconstruct and Grahn, 2007) stand out as the only episodes and geochemical disturbance were synchronous glacioeustatic sea-level Á uctuations corre- in a 210 m.y. stretch (ca. 580–370 Ma) that other- (Marshall and Middleton, 1990; Long, 1993; sponding to maximum glacial conditions wise lacks direct geologic evidence for continen- Brenchley and Marshall, 1999; Brenchley et al., associated with the end-Ordovician ice age. tal ice sheets (Hambrey and Harland, 1981). This 2003), no sedimentary succession contains The combined litho- and chemostratigraphic interval is thought to have occurred in a green- well-preserved records of all three events. house interval dominated by high atmospheric Without a À rm chronological framework, it pCO (e.g., Berner, 2006), although sedimentary is difÀ cult to make strong statements regarding †Current address: Department of Geology, Am- 2 herst College, 11 Barrett Hill Road, Amherst, Massa- records of shoreline ice in tropical Laurentia in the causes, consequences, and interrelatedness chusetts 01002, USA; [email protected] Middle to Late Cambrian time complicate this of the events.

GSA Bulletin; July/August 2011; v. 123; no. 7/8; p. 1645–1664; doi: 10.1130/B30323.1; 14 À gures; 2 tables; Data Repository item 2011133.

In this paper, we present new high-resolution enced waning effects of the thermal subsidence Tower, Homard, and Joseph Point members, and stable isotope data coupled to lithostratigraphy (Waldron et al., 1998) that began in the Ediacaran the formal Mill Bay and Schmitt Creek Mem- from the Lousy Cove and Laframboise Mem- Period around 615 Ma (Kamo and Gower, 1994). bers (Long, 2007). The Vauréal is succeeded by bers of the upper Ellis Bay Formation and the In Middle Ordovician time, the advance of the the Ellis Bay Formation, which has been divided Fox Point Member of the lower Becscie For- Taconic arc from the south and east reinvigorated into the Grindstone, Velleda, Prinsta, Lousy mation on Anticosti Island in eastern Canada Á agging subsidence rates (Waldron et al., 1998; Cove, and Laframboise Members (Fig. 2A) (Fig. 1). Because the carbon isotopic compo- Long, 2007). Back-stripping analysis using ages (Long and Copper, 1987a). The Lower Silurian sition of the global ocean changes smoothly and water depths inferred from paleontological Becscie Formation overlies the Ellis Bay; it is (rather than discontinuously) in time, carbon data suggests an increase in sedimentation rate as composed of the Fox Point and Chabot Members isotope excursions can provide time-varying the foreland basin developed through the Sand- (Copper and Long, 1989; Sami and Desrochers, signals that allow for high-resolution strati- bian and Katian Stages of the Ordovician (Long, 1992; Long, 2007). The remainder of the Anti- graphic correlation. Brenchley et al. (2003) used 2007). Sediments continued to accumulate during costi succession includes the Merrimack, Gun the Hirnantian positive carbon isotope excursion the Early Silurian, but deposition waned once River, Jupiter, and Chicotte Formations (Fig. to achieve high-resolution intrabasinal correla- Taconic thrust loading ended in the Llandovery 2B). This paper focuses on strata of the Lousy tions within Baltica, and interbasinal correlation (Long, 2007), diminishing the production of Cove, Laframboise, and Fox Point Members. between Baltica and Laurentia, improving our additional accommodation space. Thermal matu- Most workers have interpreted the Anti- understanding of the relative timing of environ- ration data, basin modeling, and offshore seismic costi Basin sedimentary record to represent a mental changes, isotopic evolution, and biotic data demonstrate that Anticosti strata were subse- subtidal, storm-influenced, carbonate ramp- events associated with the end-Ordovician ex- quently buried no deeper than 2.5–3 km and are platform (Long and Copper, 1987a; Sami and tinction and glaciation. In this paper, we employ still in the oil window (Pinet and Lavoie, 2007). Desrochers, 1992; Farley and Desrochers, 2007; the Hirnantian positive carbon isotope excur- There has been little signiÀ cant structural defor- Long, 2007; Desrochers et al., 2010). The long sion for high-resolution intrabasinal correlation mation to the island, although structural analysis axis of the island is oblique to the paleoshore- within the Anticosti Basin. By documenting indicates the presence of fracture sets produced line, with the east end occupying a more proxi- the relationships between discontinuities in the by Taconic, Acadian, and/or Jurassic events mal position and the west end occupying a more chemostratigraphic record and disconformities (Bordet et al., 2010). distal position. Previous stratigraphic work has in the lithostratigraphic record, we can construct Although over 3 km of sediment accumulated established several cycles of glacioeustatic sea- a new sequence-stratigraphic model for glacio- in the Upper Ordovician–Lower Silurian Anti- level Á uctuation in the Vauréal and Ellis Bay eustatic sea-level Á uctuations associated with costi Basin, over two-thirds of the succession is Formations, argued to have been caused by the end-Ordovician Earth history. known only from subsurface exploration (INRS- waxing and waning of continental ice sheets on Petrole, 1974). The Romaine, Mingan, Trenton– Gondwana (Desrochers et al., 2010). Geological Setting Black River, Macasty, and Princeton Lake Formations have been characterized through Chemostratigraphic Records across the The autochthonous Ordovician–Silurian Anti- borehole sampling (INRS-Petrole, 1974). The Ordovician-Silurian Boundary costi Basin was deposited on top of Grenville- lower Vauréal Formation is also conÀ ned to the age crust on the eastern margin of Laurentia subsurface, but its upper ~300 m crop out on the The global boundary stratotype section and (Waldron et al., 1998). Sedimentation in the basin surface (Long and Copper, 1987a). It has been point (GSSP) for the Hirnantian Stage is at began in the latest Cambrian as the region experi- subdivided into the informal La Vache, Easton, Wangjiawan, North China, where it coincides

Figure 1. Geologic map of Anti- costi Island, modified from English Head Desrochers and Gauthier (2009), D821 N showing locations of studied ′ sections. Approximately 900 m 49°45 of gently dipping Ordovician Laframboise Point Natiscotec Creek to Silurian stratigraphy are D801-D806, 901 920 Salmon River exposed on the island. The LEGEND D814, D816, D817, 908, 909, 911 N Laframboise Member of the ′ Chicotte Macaire Creek Jupiter D813, D815, 912 Ellis Bay Formation is the main 49°30 Lousy Cove target of this work. It is the Gun River 914 upper most member of the Ellis Merrimack Fox Pt. N

Bay Formation, occurring just ′ Becscie below the Ellis Bay–Becscie contact. Facies associations 49°15 Ellis Bay may be grouped into the west- Vauréal

Ordovician Silurian 010205 Measured ern sector (west coast and sec- section km tions near Laframboise Point: 64°30′W 64°00′W 63°30′W 63°00′W 62°30′W 62°00′W sections D801–D806, D821, 901), east-central sector (Natiscotec River tributary: section 920; Salmon River: sections D814, D816, D817, 908, 909, 911; and Macaire Creek: D813, D815, 912), and eastern sector (Lousy Cove to Fox Point: section 914).

1646 Geological Society of America Bulletin, July/August 2011 Anticosti Island chemostratigraphy and sequence stratigraphy

13 with the À rst appearance of the graptolite I Corg baseline values sit at ~–30.2‰ (Vienna Akido graptus ascensus, marking the base of the 13 Normalograptus extraordinarius (Chen et al., Peedee belemnite [VPDB]) and begin climbing Silurian System, the I Corg values have returned 13 2006). The I Corg record of a parallel section to ~–29.5‰ in the latest Katian Stage. Values to the pre-excursion baseline (Fig. 3). 180 m to the southeast of the GSSP, Wangjiawan are steady between –29.5‰ and –29.0‰ in the The GSSP for the base of the Silurian Sys- Riverside (Chen et al., 2006), hosts a prominent lower half of the Hirnantian Stage and then peak tem is at Dob’s Linn, Scotland (Holland, 1985), 13 positive I Corg excursion in Hirnantian strata at ~–28.5‰ before decreasing in the uppermost where the À rst appearance of the graptolite (Fig. 3). At this location, the pre-excursion Hirnantian beds. By the first appearance of A. ascensus provides the biostratigraphic tie

Figure 2. (A) Stratigraphy and biostratigraphy of the Ellis Bay Formation and adjoining beds, A Period Stage Stratigraphy Graptolites Conodonts Chitinozoans B modiÀ ed from Melchin (2008). Silurian Rhuddanian Fox Point A. ascensus Oulodus nathani P. nodifera Chicotte Position of the Katian-Hirnan- Bec Ancyrochitina tian boundary is contentious; 7 Laframboise Ozarkodina hassi ellisbayensis Jupiter most biostratigraphers argue ? Normalo- 6 Spinachitina for the base of the entire Ellis 5 Lousy graptus taugourdeaui Gun River Cove Bay Formation, and chemo- 4 persculptus

Lower Silurian Lower Merrimack 13 stratigraphers argue for the Prinsta Gamachig- Belonechitina Becscie

Katian 3 nathus H. base of the Laframboise Mem- Velleda Normalo- gamachiana Hirnantian Ellis Bay

Ordovician ensifer

Ellis Bay Formation Ellis Bay graptus ber. See Figure 7 for lithological 2 Grindstone extraord. legend. (B) Generalized stra- ? Vauréal

Schmitt Dicellograptus Hercochitina Katian Llandovery tigraphy of Anticosti Island, 12 100 m

Vaur Creek anceps crickmayi Upper Ordovician modi À ed from Petryk (1981); Long and Copper (1987a); Long (2007). H.—Hirnantian.

Stg. Graptolites Dob’s Linn, Scotland Laframboise Point, Wangjiawan Riverside, Anticosti Island, Canada China acuminatus/ ascensus 1 m

Rhuddannian 0.5 m 1 m persculptus

extraordinarius Hirnantian

pacificus

–30.5 –29.5 –28.5 13 Katian complexus δ Corg (‰, VPDB)

–33 –32 –31 –30 –29 –28 13 δ Corg (‰, VPDB)

041 2 3 13 δ Ccarb (‰, VPDB)

Figure 3. Carbon isotope records and graptolite biozonation from Dob’s Linn, Scotland (Silurian global stratotype section and point [GSSP]; Underwood et al., 1997), Wangjiawan Riverside (180 m from Hirnantian GSSP; Chen et al., 2006), and Laframboise Point, Anti- costi Island (current study). Locations of graptolite zones (dashed lines) are documented for Dob’s Linn and Wangjiawan, but are inferred here for Anticosti, based on matching distinctive patterns in the I13C record. The light shading identiÀ es the À rst phase of the isotope excursion (initial rise); the dark shading represents the second phase (maximum values attained). The inferred correlation of Anticosti Island stratigraphy with the GSSPs is based on the high-resolution carbon isotope stratigraphy presented here. VPDB—Vienna Peedee belemnite.

Geological Society of America Bulletin, July/August 2011 1647 Jones et al. point (Melchin and Williams, 2000). The GSSP raphy of the western end of the island is repre- acid-insoluble residue was rinsed three times is dominated by graptolitic shales, and Under- sented by a composite section assembled from in deionized water, dried in a 70 °C oven over- wood et al. (1997) documented an ~+3‰ excur- coastal outcrops measured from English Head to night, homogenized with a mortar and pestle, 13 sion in I Corg coincident with the Hirnantian Laframboise Point and Cape Henri (Fig. 4). The and weighed. Aliquots of the insoluble residue 13 13 Stage (Fig. 3). Pre-excursion I Corg values are strata of the east-central portion of the island are were used for the I Corg analyses. ~–32‰. Values in the latest Rawtheyan (end represented by outcrops along the bluffs Á anking Katian Stage) climb to ~–30.5‰. The Hirnan- the Salmon River (9 Mile Pool and 8 Mile Pool), Mass Spectrometry tian beds peak at ~–28.5‰ before decreasing. and by sections exposed in stream cuts and by The basal Silurian beds begin at the pre-excur- the road at Macaire Creek and a tributary of the Carbonate carbon isotope ratios were measured sion baseline of ~–32‰. Natiscotec River (Fig. 5). The eastern portion of at both Washington University and Harvard 13 The positive I Corg excursions at Dob’s Linn the island is represented by a composite section University. At Washington University, ~100 Rg and Wangjiawan Riverside both begin in the lat- assembled from coastal outcrops between Lousy samples of drilled carbonate powder were re- est Katian Stage, and they are both completely Cove and Fox Point (Fig. 6). Coordinates of each acted for 4 h at 72 °C with an excess of 100%

À nished by the base of the Silurian System. Fur- section are listed in Table 1. H3PO4 in He-Á ushed, sealed tubes. Evolved thermore, both excursions are characterized by In the course of stratigraphic logging, we col- CO2 was sampled with a Finnigan Gas Bench a rise to an initial elevated plateau, and then a lected and cataloged fresh À st-sized hand sam- II, and isotopic ratios were measured with a sharper spike to maximum values, followed by a ples at regular stratigraphic intervals. Samples Finnigan MAT 252 or Delta V Plus. Isotopic quick decline to pre-excursion baseline (Fig. 3). were subsequently slabbed with a rock saw. measurements were calibrated against NBS- 13 18 Thus, the Hirnantian Stage at both GSSPs is Material for I Ccarb and I Ocarb analyses was 19, NBS-20, and two in-house standards, with 13 13 characterized by enriched I Corg values, with obtained by drilling out ~100 mg of carbonate analytical errors of <0.1‰ (1X) for I Ccarb and 18 no intervening return to baseline. powder from each slab with a dental drill, target- <0.2‰ (1X) for I Ocarb. At Harvard University, The Hirnantian positive carbon isotope excur- ing À ne-grained mudstone where possible. An a VG Optima dual inlet mass spectrometer with sion has been identiÀ ed in Laurentia (Orth et al., additional slab from each sample was crushed a modiÀ ed Isocarb preparation device was used. 1986; Long, 1993; Finney et al., 1999; Kump to a homogeneous powder in a Spex 8515 Approximately 1 mg of each sample was reacted et al., 1999; Bergström et al., 2006; Melchin shatterbox with an aluminum carbide vessel. in a 90 °C common H3PO4 bath for 8–10 min. and Holmden, 2006; LaPorte et al., 2009; Young Approximately 1 g of this powder was reacted Liberated CO2 gas was collected and puriÀ ed et al., 2010), South China (Wang et al., 1993, with an excess of 6 N HCl to remove carbonate cryogenically. Sample gas and an in-house ref- 13 1997; Chen et al., 2006; Fan et al., 2009; Yan minerals in preparation for I Corg analysis. The erence gas were analyzed four times each in the et al., 2009; Zhang et al., 2009), and Baltica (Kaljo et al., 2001, 2004, 2008; Brenchley et al., 2003; Young et al., 2010). Most of these reports 13 13 have focused on either I Ccarb or I Corg, but not both, and few studies that have examined both proxies report reproducible excursions in both phases. Because of the glacioeustatic sea-level fall associated with contemporaneous Gond- wanan glaciation, Hirnantian strata are not well represented in the shallow-marine sedimentary record, and many Hirnantian sections probably contain hiatuses (Brenchley et al., 2003; Berg- A B ström et al., 2006). This fact has contributed to debate over the shape and magnitude of the carbon isotope excursion. Despite intra- and interbasinal variability, reproducible Hirnantian I13C excursions have been identiÀ ed on all three Fox Pt. Mb. Bio- paleocontinents. herm Bioherm

MATERIALS AND METHODS Bioherm Fox Pt. Mb.

Bioherm Field Work and Sample Preparation C D We logged and sampled 16 sections through the Ellis Bay and lower Becscie Formations Figure 4. Photomontage of outcrops used in composite section of the Ellis Bay Formation at during the summers of 2008 and 2009 (Fig. 1). Laframboise Point, western sector of Anticosti Island. (A) Hummocky cross-stratiÀ ed À ne We focused our high-resolution studies on the carbonate grainstones overlie nodular lime mudstones at Laframboise Point (sections D804 Laframboise Member of the Ellis Bay Forma- and 901). (B) Bioherms of the Laframboise Member are overlain by Becscie Formation tion, where previous workers (Orth et al., 1986; skeletal grainstones (D804 and 901). (C–D) Details of contact between Laframboise Member Long, 1993; Brenchley et al., 1994; Desrochers and Becscie Formation at Laframboise Point demonstrating the complex relationship be- et al., 2010; Young et al., 2010) had identiÀ ed tween the Laframboise bioherms and the overlying Becscie skeletal grainstones. Backpack 13 the Hirnantian I Ccarb excursion. The stratig- for scale in C. Visible portion of measuring stick in D is 65 cm long.

1648 Geological Society of America Bulletin, July/August 2011 Anticosti Island chemostratigraphy and sequence stratigraphy

1987a). The western sector (coastal outcrops at the western tip of the island and around Lafram- boise Point) constitutes one distinct set of Bioherm facies, dominated by carbonate rocks with very Becscie little siliciclastic input (Fig. 4). The east-central sector (outcrops at Natiscotec River tributary, Salmon River, and Macaire Creek) constitutes a Bioherm Lousy Cove second distinct suite of facies; here, the Lafram- boise Member is thinner than in the west, and siliciclastic sands are present in the lowermost A B Fox Point Member (Fig. 5). The eastern sector (coastal outcrops from Lousy Cove to Fox Oncolite Point) constitutes the third distinct suite of facies (Fig. 6). Here, the siliciclastic content is at Sandstone a maximum, and the Laframboise Member has reduced thickness. The Laframboise Member Oncolite is not well exposed in outcrop in a signiÀ cant Sandy carbonate grainstone portion of the island’s center. This leads to an uneven distribution of studied outcrops in our Lousy Cove mudstones transect across the basin (Fig. 1). Locations and names of sections are given in Table 1. Sandy carbonate Observation at outcrop and microscopic C D grainstone scales reveals excellent preservation of sedimentary structures and textures. There has been minimal recrystallization of carbonate matrix or Figure 5. Photomontage of outcrops from the east-central sector at Salmon River and Macaire skeletal material. Evidence of dolomitization is Creek. (A) Bioherms of the Laframboise Member overlying Lousy Cove Member (left bank of rare. Skeletal grains are often abundant and well Salmon River). (B) Contact between Laframboise bioherms and Becscie skeletal grainstones preserved (GSA Data Repository Fig. DR1A (right bank of Salmon River). (C) Detail of Laframboise Member, illustrating transition from and DR1D1). Bioturbation textures occur in massive sandy carbonate grainstone to oncolite platform bed, which occurs at the level of the many facies, and burrows are preserved in À ne head of the pick (section D816). (D) Contact between Lousy Cove and Laframboise Members, detail (Fig. DR1B and DR1C [see footnote 1]). 100 m upstream from section D816; bioherms are not developed here. Oncolites preserve internal laminations (Fig. DR1E [see footnote 1]).

course of each measurement. The CM-2 Cararra NBS-21 graphite, IAEA-C6 sucrose, and in- Lousy Cove Member (Ellis Bay Formation) Marble standard was measured seven times dur- house acetanilide standards. All samples were ing each run of 53 samples to calibrate samples measured in duplicate with a reproducibility In the western sector at section D804 and to the VPDB standard, provide a measure of of <0.2‰ (1X) and are reported relative to the 901 (Laframboise Point), exposure of the Lousy uncertainty, and monitor potential memory ef- VPDB scale. Cove Member begins with recessive greenish- fect associated with the common acid bath gray lime mudstones with a nodular texture system. Analytical uncertainties (1X) for both STRATIGRAPHY AND FACIES and rare centimeter-scale grainstone beds. This 13 18 I Ccarb and I Ocarb were <0.1‰, and memory facies is sharply overlain by a gray calcisiltite

effect (the contribution of any CO2 gas from The stratigraphy and biostratigraphy of the unit with 10–40-cm-thick beds and discon- the previous sample due to use of a common Ellis Bay and Becscie Formations have been the tinuous mud stringers. The calcisiltite becomes 13 acid bath) for I Ccarb was consistently <0.2‰, subject of numerous careful and detailed stud- more massive and platy up-section and develops as determined by measurement of within-run ies (e.g., Lake, 1981; McCracken and Barnes, hummocky cross-stratiÀ cation with submeter standards. A subset of samples was measured at 1981; Petryk, 1981; Nowlan, 1982; Long and wavelength in the uppermost 2 m of the mem- both laboratories as a check of interlaboratory Copper, 1987a, 1987b; Barnes, 1988; Sami and ber (Fig. 4A). The upper surface of the Lousy consistency, and no discernible difference was Desrochers, 1992; Farley and Desrochers, 2007; Cove Member is a blackened, scoured surface observed between them. Desrochers et al., 2010; Achab et al., 2010). In with ~5 cm of erosional relief. Organic carbon isotopes were measured by this section, we provide a brief outline of our In the east-central sector at sections D814– combusting tin cups containing acid-insoluble stratigraphic observations in order to provide D816 (Salmon River upstream, 9 Mile Pool) residue in a Costech ECS 4010 Elemental Ana- the context for our chemostratigraphic data. and 909 (Salmon River downstream, 8 Mile lyzer at 1000 °C. The mass of insoluble resi- The stratigraphic expression of the Ellis Bay Pool), the Lousy Cove Member begins with due combusted was varied for each sample to and Becscie Formations varies along the length greenish-gray lime mudstone with occasional give a constant peak size for CO on the mass of the outcrop belt. Facies vary laterally from 2 1 spectrometer. Evolved gas Á owed through an west to east, especially in the Ellis Bay For- GSA Data Repository item 2011133, Chemo- stratigraphy of upper Ellis Bay and lower Becscie oxidation/reduction furnace before entering mation; in the west, limestones dominate with Formations, Anticosti Island, Canada, is available a Finnegan Delta V Plus for isotopic analysis. minor shales, whereas more sandy facies occur at http://www.geosociety.org/pubs/ft2011.htm or by Isotopic measurements were calibrated against in the east (Petryk, 1981; Long and Copper, request to [email protected].

Geological Society of America Bulletin, July/August 2011 1649 Jones et al.

(Fig. 6D). This disturbed zone is composed of Fox Point Mb. shales interbedded with siltstones and calcare- Fox Point Mb. ous mudstones. Above the convoluted unit, Laframboise Mb. there are 20 cm of undisturbed beds of the same lithology. These are capped by a light-colored Laframboise Mb. resistant grainstone full of rip-up clasts of the small bioherm underlying bed. The top of this bed, which is Lousy Cove Mb. the top of the Lousy Cove Member, is clearly an erosional surface, with upright aulecerid stromatoporoid fossils truncated by the base of the A B overlying Laframboise Member oncolite.

Fox Point Mb. Laframboise Member (Ellis Bay Formation) Laframboise The Laframboise Member is deÀ ned to include the oncolite platform bed (Petryk, 1981), Lousy Cove Mb. the large distinctive bioherms, and interbiohermal limestones (Long and Copper, 1987a; Copper , 2001). In the west sector (sections D804 and 901 at Laframboise Point), the onco lite bed D is 10–20 cm thick and À lls pits and scours in the underlying Lousy Cove strata. The biohermal Laframboise Mb. unit sits on top of the oncolite bed and consists oncolite bed Fox Point Mb. of meter-scale isolated bioherms and interbiohermal limestone (Figs. 4B, 4C, and 4D). The bioherms host a complex biota; to À rst order, they contain abundant calcimicrobial textures with tabulate and rugose corals. Interbiohermal beds of the Laframboise Member are largely lime mudstone and wackestone. Laframboise Mb. C E The contact with the overlying Fox Point Member of the Becscie Formation is complex. Interbiohermal beds in the 50 cm below the level Figure 6. Photomontage of outcrops used in composite section of the Ellis Bay Formation of the bioherm tops include skeletal grainstones between Lousy Cove and Fox Point, eastern Anticosti Island (section 914). (A) Laframboise composed of brachiopod, trilobite, coral, bryo- Member (lightest strata) exposed on the east coast. (B) Detail of biohermal texture in zoan, and echinoderm material, assigned to the Laframboise Member. (C) Detail of oncolitic texture within Laframboise Member. Fox Point Member of the Becscie Formation. In (D) Massive Laframboise Member between disturbed beds of upper Lousy Cove Member some places, these grainstones appear to truncate (below) and well-bedded grainstones of Becscie Formation (above). (E) Erosional contact the bioherms, but in other places, the bioherms between overlying Becscie Formation skeletal grainstones and underlying oncolite platform can be seen to be growing out over discon- bed of Laframboise Member. tinuous grainstone beds (Fig. 4D); this complicated interÀ ngering geometry suggests that the Laframboise Member bioherms were still grow- silt and hummocky cross-stratiÀ cation. This unit In the east sector at section 914 (Lousy Cove, ing, perhaps sluggishly, when the À rst incursion is overlain by several meters of rubbly weather- Fig. 6A), the bottom exposure of the Lousy Cove of Becscie skeletal grainstones was deposited. ing concretionary silty limestone. Two meters of Member is a cross-bedded calcarenite bed with In the east-central sector, the oncolite plat- mudstone separate the concretionary layer from 10–20-cm-thick foreset beds. Intraclasts are form bed of the Laframboise Member is 30– a 4-m-thick unit composed of very dark-gray pervasive, and the top of the unit is burrowed, 40 cm thick and contains abundant oncoids, planar laminated limestone with Chondrites, with a well-rounded quartz pebble lag. This corals, and stromatoporoids (Fig. 5C). It has an interbedded with gray brachiopod-bearing lime is succeeded by a bioturbated nodular muddy erosional contact with the underlying massive mudstone. Near the top, the member is com- wackestone with corals and brachiopods. Three sandy grainstone unit at the top of the Lousy posed of thin calcareous shales interbedded with meters of shales interbedded with calcareous Cove Member. The oncolite bed grades into both planar and cross-stratiÀ ed sandy grainstone mudstones and carbonate grainstones lie above overlying bioherms where they are present. The beds (Figs. 5A and 5D). The uppermost unit is the nodular layer. There are frequent skeletal bioherms are not as closely spaced in the east- a 50-cm-thick, massive buff-colored sandy car- (often crinoid) packstone and grainstone inter- central sector compared as to the west, and some bonate grainstone with a pitted and scoured top beds, some of which form gutters. Twenty centi- reach thicknesses of 8 m. In exposures where surface (Fig. 5C). At section 912 at Macaire meters of lime grainstones with hummocky bioherms are absent, the oncolite bed constitutes Creek, mudstone rip-up clasts are found in the cross-stratiÀ cation separate the shales from a the entirety of the Laframboise Member (Figs. massive sandy grainstone. 1-m-thick zone of intense internal deformation 5C and 5D).

1650 Geological Society of America Bulletin, July/August 2011 Anticosti Island chemostratigraphy and sequence stratigraphy

TABLE 1. LOCATIONS OF MEASURED SECTIONS USED IN THE STUDY CHEMOSTRATIGRAPHIC RESULTS Anticosti Island Section Locations Section ID Latitude (°N) Longitude (°W) Location name Stable isotopic data are tabulated in GSA Western sector Data Repository Table DR1 (see footnote 1). D821 49°54.072′ 64°29.458′ English Head D801 49°49.567′ 64°26.574′ Junction Cliff D802 49°48.855′ 64°25.781′ Western Anticosti Island D803 49°48.646′ 64°25.520′ D804, 901 49°48.350′ 64°25.259′ Laframboise Point D805 49°47.763′ 64°22.735′ Cape Henry At Laframboise Point in the west (Fig. 7), D806 49°47.545′ 64°23.066′ Cape Henry Lousy Cove Member mudstones record a base- 13 East-central sector line I Ccarb value of ~0.5‰. An isotopic dis- 920 49°29.530′ 62°32.631′ Natiscotec tributary continuity separates the top of the Lousy Cove 911 49°23.853′ 62°23.583′ 9 Mile Pool, right bank D816, D817, 908 49°24.015′ 62°23.175′ 9 Mile Pool, left bank mudstones from the overlying sandy carbon- ′ ′ 13 D814, 909A/B 49°24.328 62°21.420 8 Mile Pool ate grainstones. In the coarser unit, I Ccarb and ′ ′ 912, 912A 49°22.762 62°12.320 Macaire Creek Quarry I13C rise by 2‰. Although there is strong D813 49°22.838′ 62°12.230′ Macaire Creek stream A org D815 49°22.498′ 62°12.225′ Macaire Creek stream B À eld evidence for an erosional surface separat- Eastern sector ing the grainstones from the onco lite platform 13 914/A/B 49°19.198′ 61°51.399′ Lousy Cove bed, the continuity of I Ccarb across this surface indicates that not much time is missing 13 at this boundary. Above the oncolites, I Ccarb climbs to +4‰ and fluctuates around that value through the biohermal unit. Similarly, 13 In the east sector at section 914 (Lousy Cove), 5 to 10 cm of relief on the underlying Lafram- I Corg shows a parallel rise below the onco- the oncolite platform bed of the Laframboise boise Member (Fig. 6E). Dark-gray skeletal lite bed and Á uctuations around the maximum Member is 50–60 cm thick and contains abun- grainstones at the bottom of the member are above it (Fig. 7). Up section, high-resolution dant corals and oncolites (Fig. 6A). The base replaced with hummocky cross-stratiÀ ed À ne sampling shows that there is a smooth decline 13 truncates aulecerid stromatoporoid fossils of the calcarenite within 2 m of the base. Normally in I Ccarb from the +4‰ maximum down to underlying Lousy Cove Member. Oncolites are graded bioclastic interbeds are common. +0.5‰ from the top of the biohermal unit nucleated around clasts of the underlying bed (Fig. 6C). Bioherms are less abundant here than in the other sectors, and they are considerably smaller (Fig. 6B).

Fox Point Member (Becscie Formation) 20 The base of the Becscie Formation at sections D804 and 901 (Laframboise Point) in the western sector is characterized by 1 m of interbedded grainstones and lime mudstones (Figs. 4B, 4C, and 4D) that appear in some places 15 to interÀ nger with the Laframboise Member bioherms (Fig. 4D). Some grainstones display hummocky cross-stratiÀ cation or wave ripples. A few meters up-section, the grainstone beds become thinner and less common, and mud- 10 stone dominates. Height in section (m) In the east-central sector, the Becscie For- mation begins with thin resistant sandstone that overlies the Laframboise oncolite platform Lousy CoveLousy Laframboise Point Fox bed where bioherms are absent (Fig. 5D). The 5 sandstone frequently displays low-angle cross- stratiÀ cation. Above the sandstone, the Fox Point Member typically becomes recessive and poorly exposed for 2 to 4 m. At section 911, on 0 the right bank of the Salmon River, across from –5–4 –3 –2 –1024135 –30 –28 –26 –24 –22 2628 30 32 34 18 13 13 13 the upstream section (9 Mile Pool), bioherms of δ Ocarb, δ Ccarb (‰, VPDB) δ Corg (‰, VPDB) Δ C (‰, VPDB) the Laframboise Member are directly overlain with skeletal grainstones with a diverse fauna, Figure 7. Geochemistry and lithostratigraphy of sections D804 and 901 from Laframboise including ramose bryozoans and nautiloids. Point, western Anticosti Island. Oncolite platform bed data are shaded in gray. Filled data The base of the Fox Point Member at section points are from Ellis Bay Formation; open data points are from Becscie Formation. See 914 (Lousy Cove) in the eastern sector displays Figure 8 for legend. VPDB—Vienna Peedee belemnite.

Geological Society of America Bulletin, July/August 2011 1651 Jones et al. through the lowest 2 m of the Becscie skeletal 40 13 grainstones. The I Corg values are observed to decline in parallel.

East-Central Anticosti Island at Salmon River and Macaire Creek

no 909B

Fox Point Mb. Point Fox exposure At Salmon River, in the east-central sector of the island, the Laframboise Member is exposed LFB on the riverbanks for over a kilometer. At the 30 downstream section (8 Mile Pool, section 909, Fig. 8), the member is <1 m thick and lacks 909A 13 bioherms. The I Ccarb pattern in the underlying Lousy Cove Member has a well-deÀ ned base- –30 –28 –26 –24 –22 26 28 30 32 13 13 δ13 Δ13 line of +0.5‰. Both I Ccarb and I Corg begin to Corg (‰, VPDB) C (‰, VPDB) climb 2 m below the Lousy Cove–Laframboise contact. There is a ~1‰ step up in both proxies 32 across the base of the oncolite platform bed, above which 13C continues to rise to +4‰. Mb. Pt. Fox I carb 20 LFB The cross-stratiÀ ed sandstone bed at the base 30 of the Fox Point Member hosts a 2‰ downturn Mb. Cove Lousy 13 18 in I Ccarb (and a simultaneous drop in I Ocarb), Height in section (m) Height in section (m) which may be a diagenetic artifact due to high Mb. Cove Lousy 28 cement content of the bed. Above the cross- 012 345 13 13 B δ Ccarb (‰, VPDB) stratiÀ ed bed, I Ccarb rises again to a peak of +4.5‰, before declining to +3‰ at the top of the D814 LEGEND 13 exposure. In this interval, I Corg moves in the δ18 δ13 13 13 13 O C same direction as I Ccarb, but ) C (= I Ccarb – Disconformity 13 10 I Corg) steps up to ~29‰. The succeeding 2 m Lime mudstone of stratigraphy are not exposed. Five meters of Concretionary wackestone the Fox Point Member are exposed at the top Plane-laminated wackestone 13 Cross-laminated sandy of the outcrop, and this interval shows I Ccarb grainstone values constant at +1.8‰, i.e., signiÀ cantly en- Skeletal grainstone riched relative to the pre-excursion values in the Nodular mudstone Oncolitic grainstone

Lousy Cove Member. It is also enriched relative Mb.? Prinsta Coral-microbial bioherm to the lithologically equivalent strata at Lafram- Massive carbonaceous sandstone boise Point at the western end of Anticosti Cross-bedded carbonaceous Island . Throughout the I13C -enriched beds, 0 sandstone carb -8–6–4 –2 0 2 4 6 Interbedded calcareous I13C values monotonically increase such that A 18 13 shale and limestone org δ Ocarb, δ Ccarb (‰, VPDB) )13C returns to its pre-excursion value of 28‰ at the top of the section. Figure 8. (A) Geochemistry and lithostratigraphy of the composite section from the Salmon Below the Laframboise Member, at the down- River downstream section, central Anticosti Island. Data are from sections D814 and 909A/B. stream locality (8 Mile Pool, section D814, (B) Detailed stratigraphic column of section 909A/B at the Ellis Bay–Becscie Formation Fig. 8), I13C values are relatively invariant 13 carb contact. Thin oncolite platform bed data are shaded in gray. Pre-Hirnantian I Ccarb excur- at ~0‰, with one signiÀ cant exception; a posi- sion is shaded in gray at the Prinsta–Lousy Cove boundary. Filled data points are from Ellis tive excursion to +1.5‰ occurs at a lithological Bay Formation; open data points are from Becscie Formation. LFB—Laframboise Mem- change ~20 m below the Laframboise Member ber. VPDB—Vienna Peedee belemnite. that may correspond to the Prinsta–Lousy Cove boundary. Approximately 15 m of Lousy Cove 13 stratigraphy with an isotopic composition close ceous sandstones, I Ccarb reaches a peak of +2‰ in the lowermost Becscie Formation fol- 13 to 0‰ separate this excursion from the larger +4.1‰ (Fig. 9A). There are two I Ccarb discon- lows these enriched values. The Becscie Forma- 13 I Ccarb excursion described previously (Fig. 8). tinuities: one across the surface separating the tion remains isotopically enriched over the next Two kilometers upstream (9 Mile Pool, sec- Lousy Cove mudstones from the Lousy Cove 10 m of stratigraphy, declining over that interval tions D816, 908, and 911; Fig. 9), bioherms are sandy grainstones, and another across the Lousy until it reaches 0‰ at 20 m above the top of the variably present in the Laframboise Member Cove–Laframboise contact. Measurements Laframboise Member. (Fig. 5). In the mudstones of the upper Lousy from the top of a Laframboise bioherm into the Ten kilometers east of Salmon River at 13 Cove Member, I Ccarb maintains a baseline overlying skeletal grainstones of the Becscie Macaire Creek (sections D815 and 912; Fig. 13 value of 0.7‰. Isotope values in the uppermost Formation (section 911; Fig. 9B) show that the 10), I Ccarb values jump discontinuously across 13 Lousy Cove grainstones increase to +2‰, and uppermost biohermal carbonates have I Ccarb the surface separating the Lousy Cove mud- in the overlying Fox Point Member carbona- of ~+4.5‰. A step-function drop to just under stones from the overlying Lousy Cove sandy

1652 Geological Society of America Bulletin, July/August 2011 Anticosti Island chemostratigraphy and sequence stratigraphy

15 D817 Fox Point Member Point Fox 10 Height in section (m) no

5 exposure

LFB D816

0 Mb. LC –4 –2 024–32 –30 –28 –26 –24 –22 22 24 26 28 30 32 δ18 δ13 δ13 Δ13 A Ocarb, Ccarb (‰, VPDB) Corg (‰, VPDB) C (‰, VPDB)

0 911 LFB Pt. Fox Height in section (m) –4 –2 024–32 –30 –28 –26 –24 –22 22 24 26 28 30 32 18 13 13 13 B δ Ocarb, δ Ccarb (‰, VPDB) δ Corg (‰, VPDB) Δ C (‰, VPDB)

Figure 9. (A) Geochemistry and lithostratigraphy of the composite section from the left bank of the Salmon River upstream section, central Anticosti Island. Data are from sections D816 and D817. Section D816 was measured between bioherms. (B) Geochemistry and lithostratigraphy of the section from the right bank of the Salmon River upstream section, central Anticosti Island. Data are from section 911, through uppermost exposure of partially exposed bioherm and overlying Becscie strata. Oncolite platform bed data are shaded in gray. Filled data points are from Ellis Bay Formation; open data points are from Becscie Formation. LC—Lousy Cove, LFB—Laframboise Member, VPDB—Vienna Peedee belemnite. See Figure 8 for legend.

grainstones. Isotope values within the sandstone across the lower contact with the Lousy Cove than 90 °C (McCracken and Barnes, 1981). The rise from 2‰ to 4‰ and then decline back to- mudstones (Fig. 11). The transition to the Fox structure of the basin is simple. Strata across the ward 2‰. The lower Becscie Formation local Point Member of the Becscie Formation does island have a constant dip of ~2° to the southwest minimum observed within the cross-stratiÀ ed not include an immediate return to baseline (Fig. 1), and few faults are exposed, indicating 13 grainstones at Salmon River (Fig. 8) is not re- I Ccarb values. The lowest 4 m of the Becscie that the rocks of the island have experienced produced at Macaire Creek. Organic carbon Formation on the east coast are enriched by 2‰ minimal deformation. Microscopic analysis of isotope values decline above the oncolite bed, relative to the west. polished slabs (Fig. DR1 [see footnote 1]) and 13 in parallel with I Ccarb; however, they do so at thin sections conÀ rms observations made at the a different rate, producing rising )13C values. DO ANTICOSTI ISLAND STABLE outcrop scale, i.e., that there has been little ob- The lower Becscie Formation is poorly exposed ISOTOPE RECORDS REFLECT A servable recrystallization. The relatively pristine above the sandstone at this outcrop, but the few PRIMARY SEAWATER SIGNAL? state of the Anticosti carbonate rocks indicates measured samples in the lowermost 2 m regis- that they have not been subject to extensive dia- 13 ter I Ccarb of ~+2‰, i.e., signiÀ cantly elevated Diagenetic Considerations genetic alteration. from pre-excursion baseline. Carbonate rocks that have been isotopically The carbonate rocks exposed on Anticosti altered by meteoric diagenesis are frequently 18 Eastern Anticosti Island Island are preserved exceptionally well. An depleted in I Ocarb because meteoric waters analysis of the conodont color alteration index tend to have very low oxygen isotope ratios. The Laframboise Member is <1 m thick at (CAI) of specimens from the Ellis Bay and Because diagenetic Á uids contain little carbon exposures on the east coast of the island (Fig. 6). Becscie Formations shows that most conodont but abundant oxygen, carbonate rocks that It is largely oncolitic from the base to the top. elements have a CAI of 1.0–1.5, correspond- have experienced extensive diagenetic over- 13 There is a large (>3‰) discontinuity in I Ccarb ing to maximum burial temperatures of less printing can maintain original carbon isotope

Geological Society of America Bulletin, July/August 2011 1653 Jones et al.

Figure 10. Geochemistry and 6 lithostratigraphy of the com- no exposure posite section from the Macaire Creek section, central Anti costi Island. Data are from sections 4 D815 and 912. Oncolite plat- LFB form bed data are shaded in gray. Filled data points are

from Ellis Bay Formation; open Height in section (m) data points are from Becscie 2 Formation. LFB—Laframboise

Member, VPDB—Vienna Peedee D815 912 Lousy Cove Mb. Cove Lousy Member Point Fox belemnite. See Figure 8 for legend. 0 –6–4 –2 024–32–30 –28 –26 –24 –22 24 26 28 30 32 34 18 13 13 13 δ Ocarb, δ Ccarb (‰, VPDB) δ Corg (‰, VPDB) Δ Corg (‰, VPDB)

8 Figure 11. Geochemistry and lithostratigraphy of the composite section from Lousy Cove 6 to Fox Point Members, eastern LFB Anticosti Island. Data are from section 914. Oncolite platform 914 bed data are shaded in gray. 4 Filled data points are from

Height in section (m) Ellis Bay Formation; open data points are from Becscie Forma- tion. LFB—Laframboise Mem- 2 ber, VPDB—Vienna Peedee

Lousy Cove Member Cove Lousy Member Point Fox belemnite. See Figure 8 for legend. 0 –8–6–4–20246–30 –28 –26 –24 –22 24 26 28 30 18 13 13 13 δ Ocarb, δ Ccarb (‰, VPDB) δ Corg (‰, VPDB) Δ Corg (‰, VPDB)

ratios despite exhibiting very depleted oxygen recorded in the Anticosti Basin, and cathodo- Lateral I13C Gradients in the isotope ratios (Banner and Hanson, 1990). luminescence studies have suggested some Anticosti Basin? However, any small amount of alteration of degree of meteoric alteration. However, exam- I13C during meteoric diagenesis would likely ination of the isotopic record of this interval Lateral gradients in I13C of up to 4‰ have be derived through reaction with respired car- at Laframboise Point (the section for which been documented in modern seawater and sedi- bon from decomposed organic matter. This the most data are available; Fig. 7) shows that ment of Florida Bay and the Bahama Banks respired carbon would have a highly depleted Laframboise strata are enriched, not depleted, (Patterson and Walter, 1994), late Cenozoic 13 13 I C composition, thus depleting the I Ccarb in both oxygen and carbon isotope ratios. This sediments of the Bahamas (Swart and Eberli, signature of the meteorically altered carbon- result rules out the possibility of meteoric al- 2005), and Paleozoic sedimentary rocks of the ate rocks. We can test the Anticosti Island teration of the isotopic record. North American midcontinent (Holmden et al., samples for this diagenetic effect by examin- The low conodont alteration index and lack 1998), all associated with carbonate platform ing the rocks deposited in the shallowest water of evidence for meteoric diagenesis suggest that and epeiric sea settings. These gradients can de- 13 depths, which would therefore be the rocks I Ccarb values measured in Anticosti limestones velop when platform waters are not efÀ ciently most readily exposed to meteoric diagenesis. likely reÁ ect the I13C composition of the sea- mixed with the open ocean, allowing a geo- As discussed already, the Laframboise Mem- water from which they precipitated (Long, 1993; chemical evolution distinct from that of the open ber represents the shallowest water depth Desrochers et al., 2010; Young et al., 2010). ocean. Depletion of I13C due to the oxidation of

1654 Geological Society of America Bulletin, July/August 2011 Anticosti Island chemostratigraphy and sequence stratigraphy isotopically light marine organic matter can oc- New and existing isotopic data for the Anti- of the discontinuities and imply asynchronous cur during the “aging” of poorly mixed water on costi strata allow for basin-scale comparisons of deposition of the strata of the Laframboise and 13 the platform top (Patterson and Walter, 1994). I Ccarb gradients before and during the Hirnan- Fox Point Members across Anticosti Island 13 13 In order to test whether a lateral I C gradi- tian positive excursion. These I Ccarb data do (Fig. 12); the stratigraphic expression of the ent existed in the Anticosti Basin, it is necessary not support the presence of a lateral isotopic I13C excursion is different at the western end, 13 to compare the I Ccarb values of coeval samples gradient in the Anticosti Basin during the depo- the east-central sector, and the eastern end, as from across the basin. We perform this test À rst sition of the Ellis Bay Formation. discussed next. with strata hosting the positive carbon isotope 18 excursion and second with older strata lower in I Ocarb as a Proxy for Climate Variables? Base of the Hirnantian Stage on the section. Anticosti Island 13 18 The peak of the Laframboise Member I Ccarb The potential for I Ocarb to provide detailed excursion is of the same absolute value in all stratigraphic records of paleoclimate has long The stratigraphic extent of the Hirnantian 13 parts of the island; the maximum I Ccarb value been recognized (Urey, 1947). Studies of the Stage on Anticosti Island has been a conten- measured in each of the three geographic sec- isotopic composition of foraminifera in Pleisto- tious issue, with biostratigraphers often arguing tors is 4.5‰. If a lateral gradient had existed, cene deep-sea sediments elucidated the roles that the entire Ellis Bay Formation is Hirnantian the expectation would be for maximum values that temperature (Emiliani, 1966) and ice vol- and chemostratigraphers arguing that only the of the excursion to vary as a function of distance ume (Shackleton, 1967) have played in the upper most Lousy Cove and Laframboise Mem- 18 from the shoreline (Immenhauser et al., 2003; I Ocarb record of marine carbonate sediments. bers are Hirnantian. Copper (2001) and Jin and Melchin and Holmden, 2006). The maximum In studies of Paleozoic carbonate rocks, signiÀ - Copper (2008) considered the entire Ellis Bay 18 value obtained in the western sector is 4.6‰ cant advances have been made using the I Ocarb Formation to be Hirnantian, based on their dis- at section 901; the maximum value obtained in of low-Mg calcite shells, which are more resis- covery of Hindella and Eospirigerina brachio- the east-central sector is 4.5‰ at section 911; tant to diagenetic overprinting than other skel- pods at the bottom of the Grindstone Member the maximum value obtained in the east sector etal carbonate mineralogies (Veizer et al., 1999). and Hirnantia within the Prinsta Member. How- is 4.4‰ at section 914. The fact that this maxi- In the present study, we focus not on brachiopod ever, as reviewed in Kaljo et al. (2008), Hindella mum value is invariant between each of the three shell material, but on À ne-grained micritic com- and Eospirigerina have their À rst appearances in geographic sectors of the basin is strong evi- ponents that provide a stratigraphically continu- Baltic sections in the Katian Stage and therefore dence that no basin-scale lateral carbon isotope ous sampling opportunity. For this reason, we may not be diagnostic of the Hirnantian Stage. 18 gradient existed during the deposition of the do not emphasize interpretations of the I Ocarb Melchin (2008) recently identiÀ ed graptolites Laframboise Member. Similarly, the baseline data. It should be noted that in the western sec- indicative of the Hirnantian Normalograptus 13 18 I Ccarb values of the pre-excursion Lousy Cove tor, I Ocarb shows a positive excursion coinci- persculptus zone in the Lousy Cove Member. If strata are consistent across the island, falling in a dent with the I13C excursion in the Laframboise these identiÀ cations are correct, (1) the base of narrow range between 0.5‰ and 0.7‰, and ex- Member, as would be expected due to changes the Hirnantian is below the Lousy Cove Mem- 18 hibiting no systematic east-west variation. This in temperature and I Oseawater associated with ber, and (2) the positive carbon isotope excur- 18 suggests that no lateral carbon isotopic gradient glaciation (Fig. 7). However, this I Ocarb ex- sion in the Laframboise Member begins in the existed in the time immediately before the start cursion is not reliably reproduced in the other N. persculptus zone, and not in the N. extraor- of the Hirnantian excursion. measured sections. dinarius zone (Melchin and Holmden, 2006; For strata older than the top of the Lousy Melchin, 2008), as argued by Baltic chemo- Cove Member, we appeal to previously pub- COMPLETENESS AND stratigraphers (Brenchley et al., 2003; Kaljo 13 lished I Ccarb data tied to basinwide correla- STRATIGRAPHIC RANGE OF et al., 2008). However, there has not been wide- tions of transgressive-regressive (TR) cycles of THE HIRNANTIAN POSITIVE spread agreement on the interpretation of the Desrochers et al. (2010) (Fig. DR2 [see foot- ISOTOPE EXCURSION graptolite record (Delabroye and Vecoli, 2010). 13 note 1]). Desrochers et al. (2010) identiÀ ed Melchin (2008) suggested that a small I Ccarb À ve TR cycles in the upper Vauréal and Ellis Previous chemostratigraphic studies of the peak in the Velleda and Grindstone Members Bay Formations, which they traced across the Hirnantian Stage on Anticosti Island and in the at the western exposures of the Ellis Bay For- length of the Anticosti Basin and interpreted Baltics have concluded that Anticosti Island does mation (Long, 1993) may correspond to the as 400 k.y. eccentricity-forced Milankovitch not preserve the entirety of the Hirnantian positive initial rise in the I13C record of the Hirnantian 13 cycles. Because these TR cycles are thought to carbon isotope excursion (e.g., Brenchley et al., excursion, and the I Ccarb peak in the Lafram- be a response to glacioeustatic sea-level Á uctu- 2003; Bergström et al., 2006). Comparing the boise Member may correlate with the second ations, they should be isochronous. Therefore, original I13C data of Long (1993) with their data rise in the Hirnantian excursion, a viewpoint boundaries between the TR cycles can serve from Baltica, Brenchley et al. (2003) suggested also expressed by Desrochers et al. (2010) and 13 13 as time lines, and the I Ccarb values at those that most of the I C excursion was missing Achab et al. (2010). In this framework, the en- boundaries can be compared in order to test for due to a signiÀ cant stratigraphic gap within the tire Ellis Bay Formation represents the most the existence of lateral isotopic gradients. The Laframboise Member. Bergström et al. (2006) expanded section of Hirnantian stratigraphy in data are variable due to low sampling resolu- and Desrochers et al. (2010) reached a similar the world, with tens of meters of strata having 13 13 tion and difÀ culties associated with precise conclusion using enhanced data sets from Anti- I Ccarb of 0‰ separating two Hirnantian I Ccarb selection of maximum regressive surfaces in costi Island and argued that two hiatuses exist— maxima. The underlying Katian and overlying conformable successions, but there is no sys- one at the base of the oncolite platform bed, and Llandovery sections on Anticosti Island are 13 tematic gradient to the I Ccarb values from one one at the Ellis Bay–Becscie contact. As de- in fact extremely thick. However, there are no end of the island to the other (see supplemen- scribed in the following, our new chemostrati- documented Hirnantian records anywhere else tary information [see footnote 1]). graphic data reÀ ne the timing and magnitudes in the world for which I13C returns to baseline in

Geological Society of America Bulletin, July/August 2011 1655 Jones et al.

Figure 12. East-west geographic 13 I Ccarb gradients in key litho- Base of oncolite platform bed Base of Becscie skeletal grainstones logical transitions of the upper 4.0 4.0 Ellis Bay and lower Becscie 13 A B formations. (A) I Ccarb values of basal oncolite platform bed 3.0 3.0 (Laframboise Member) increase from west to east on the rising carb carb

13 C 2.0 C 2.0 limb of the Hirnantian I Ccarb 13 13 excursion, indicating that the δ δ oncolite platform bed is time 13 1.0 1.0 transgressive. (B) I Ccarb values of À rst skeletal grainstones of the Fox Point Member (Becscie 0.0 0.0 Formation) decrease from east 901 920 914 912 901 912 914 911 909A to west on the descending limb D804 D816 909B D804 of the Hirnantian I13C excur- Lafram. Natis. Salmon Macaire Lousy Lafram. Salmon Macaire Lousy carb Point Creek River Creek Cove sion, indicating that this unit of Point River Creek Cove the Fox Point Member is time WestEast-Central East WestEast-Central East transgressive.

the middle of the excursion, which would be the Hirnantian brachiopod fauna and hosts the start zoan biostratigraphy can accurately identify case on Anticosti Island if the entire Ellis Bay of the Hirnantian carbon isotope excursion and the Hirnantian Stage on Anticosti Island. 13 Formation were Hirnantian. is equivalent to the S. taugourdeaui zone. The shape of the I Ccarb excursion within Recently, Achab et al. (2010) presented new As pointed out by Achab et al. (2010, the Laframboise Member at Laframboise Point chitinozoan records from Anticosti and used p. 193), “the age of the B. gamachiana zone matches the architecture of the records from the them to argue for a thick Hirnantian section ex- is...difÀ cult to establish and can only be in- GSSPs at Dob’s Linn and Wangjiawan River- tending down to the base of the Ellis Bay For- ferred.” This inference was achieved by not- side (Fig. 3). Although previous workers did not mation on the west coast. Achab et al. (2010) ing that the chitinozoan species Tanuchitina À nd this to be a compelling match (Fan et al., documented the succession of four chitinozoan laurentiana has been found in association with 2009; Melchin and Holmden, 2006), they biozones throughout Anticosti strata—Herco- both B. gamachiana in the middle Ellis Bay were working with the limited data set of Long chitina crickmayi, Belonechitina gamachiana, Formation and with H. crickmayi in the upper (1993). The higher-resolution data presented Spinachitina taugourdeaui, and Ancyrochitina Vauréal Formation (which is demonstrably of here makes the comparison clearer. In the west- ellisbayensis. The H. crickmayi zone is present Katian age). The range of T. laurentiana in ern sector at Laframboise Point (sections D804 throughout the upper Vauréal Formation. On the the Vinnini Formation (Nevada) begins in the and 901; Fig. 7), the excursion begins below the west coast, the succeeding B. gamachiana zone Katian Dicellograptus ornatus zone, where it oncolite platform bed, in the uppermost beds extends from the bottom of the Ellis Bay Forma- is observed to occur with H. crickmayi; it ex- of the Lousy Cove Member (Facies 6 and 7 of 13 tion through a level 1 m below the top of the bio- tends into the Hirnantian N. extraordi narius Desrochers et al., 2010). In these beds, I Ccarb turbated mudstones of the Lousy Cove Member. zone. B. gamachiana is absent from the Vin- shifts from pre-excursion baseline to ~+2‰. A The remainder of the Lousy Cove Member be- nini Formation (as it is from the rest of Lauren- discontinuous shift upward occurs across the longs to the S. taugourdeaui zone. The Lafram- tia). Because the Ellis Bay samples that contain contact with the oncolite platform bed at the boise Member is devoid of chitinozoans, but the T. laurentiana lack H. crickmayi, Achab et al. base of the Laframboise Member, after which 13 overlying Fox Point Member contains chitino- (2010) suggested that the middle Ellis Bay For- I Ccarb ascends steeply to its maximum values. zoans indicative of the A. ellisbayensis zone. mation could be of early Hirnantian age. This The excursion returns to baseline across a thin The chitinozoan biozones vary in their util- argument requires the absence of B. gamachi- portion of the stratigraphy at the Laframboise– ity for global correlation among carbonate- ana in the Vinnini Formation to be insigniÀ - Fox Point transition. This overall pattern is the dominated successions. Achab et al. (2010) cant, but it requires the absence of H. crickmayi same as that seen in the complete Hirnantian correlated the H. crickmayi biozone with the in the Ellis Bay Formation to be highly signiÀ - sections in China and Scotland (Fig. 3). We Dicellograptus anceps graptolite zone, which cant. Furthermore, even if the B. gamachiana therefore place the base of the Hirnantian Stage is generally accepted to be of Katian age. zone on Anticosti could be correlated with at Laframboise Point at the transition within the B. gamachiana occurs in the Katian-aged conÀ dence to the range of T. laurentiana above Lousy Cove Member between the nodular mud- upper Pirgu Stage of Estonia (Brenchley et al., the D. ornatus graptolite zone in Nevada, that stones and the overlying grainstones and calci- 2003; Kaljo et al., 2008), directly underly- still would leave the Katian-aged PaciÀ cograp- siltites, ~3 m below the oncolite platform bed. 13 ing the Hirnantian-aged Porkuni Stage. This tus paciÀ cus graptolite zone as a viable cor- There are additional positive I Ccarb peaks 13 argues for a Katian age for the B. gamachi- relative. For these reasons, we do not À nd the lower in the stratigraphy. A small I Ccarb peak ana zone (Brenchley et al., 2003; Kaljo et al., argument assigning the B. gamachiana zone (+~2‰) is recorded ~20 m below the Lafram- 2008; but see arguments of Achab et al., 2010). to the Hirnantian Stage to be compelling, and boise Member in the east-central sector (Salmon The Porkuni Stage contains an unambiguous therefore do not believe that existing chitino- River downstream, section D814; Fig. 8) and

1656 Geological Society of America Bulletin, July/August 2011 Anticosti Island chemostratigraphy and sequence stratigraphy

13 ~50 m below the Laframboise Member in the ~2.9‰). In the eastern sector, the oncolite con- The I Ccarb chemostratigraphy suggests that the western sector (Junction Cliff, section D801; stitutes the entirety of the Laframboise Member, Ellis Bay–Becscie contact is time transgressive 13 Fig. 13). As demonstrated by Desrochers et al. and I Ccarb begins at 3.7‰. Brenchley et al. across the island; the lithostratigraphic bound- (2010) and Achab et al. (2010) and conÀ rmed (2003) pointed out that the Hirnantian posi- ary should not be taken as a chrono stratigraphic here, lithofacies are diachronous across the tive carbon isotope excursion can be used as a boundary. Because diagnostic graptolites, cono- island, making it conceivable that these two “ruler” against which the rapid environmental donts, and chitinozoans cannot precisely pin +2‰ excursions are the same event. We suggest changes of the Late Ordo vician can be ordered. the base of the Silurian on Anticosti Island, and that this excursion is Katian (pre-Hirnantian) and Brenchley et al. (2003) used the excursion to because lithofacies are diachronous across the 13 may correlate with the pre-Hirnantian Paroveja achieve high-resolution stratigraphic correla- island, we suggest the use of the I Ccarb curve excursion in Baltica (Ainsaar et al., 2010). tions between Laurentia and Baltica; we use it to as an aid for identifying the Ordovician-Silurian There is chemostratigraphic evidence for the achieve high-resolution basin-scale correlations boundary. The isotope curves from the Hirnan- entire Hirnantian I13C excursion in the upper- within the Anticosti Basin. Using the ascending tian and Silurian GSSPs (Fig. 3) show that fol- 13 most Lousy Cove Member and Laframboise limb of the excursion as a time-varying signal, lowing the Hirnantian excursion, I Corg returned Member, extending into the lower Fox Point we infer that deposition of the oncolite platform to baseline at the Ordovician-Silurian boundary. 13 Member, and at least one separate I Ccarb peak bed was diachronous, progressing from west to Therefore, chemostratigraphic correlation sug- lower in the pre-Hirnantian portion of the Ellis east, as would be expected on a west-dipping gests that the Ordovician-Silurian boundary on Bay Formation. Comparing these data with the platform during a transgression. Anticosti Island is located in the lower Becscie global I13C database for the Upper Ordovician, The character of the deposits directly under- Formation (Achab et al., 2010). Our results we conclude that the base of the Ellis Bay For- lying the oncolite platform beds is markedly dif- demonstrate that this chemostratigraphic marker mation should be placed in the Katian Stage, ferent in the eastern sector compared to the rest occurs at different levels within the Becscie For- perhaps near the K2-K3 boundary (Bergström of the island. There are no sub-oncolite grain- mation and is dependent upon geography; the et al., 2009, 2010; Ainsaar et al., 2010). This stones in the uppermost Lousy Cove Member of Ordovician-Silurian boundary is 1–2 m above interpretation is at odds with interpretations of the eastern sector. The mudstones that directly the Ellis Bay–Becscie contact in the west and 13 Anticosti Island stratigraphy that place the base underlie the oncolite have I Ccarb values of ~20 m above the contact at Salmon River. We of the Hirnantian Stage at the bottom of the ~0.5‰, whereas the sub-oncolitic grainstones in did not recover a return to isotopic baseline in 13 Ellis Bay Formation based on the brachiopod the other sectors have I Ccarb values of ~2.0‰. our sampling of the eastern sector. (Copper , 2001; Jin and Copper, 2008), grapto- These observations suggest that the uppermost lite (Melchin, 2008), and chitinozoan (Achab beds of the Lousy Cove Member in the eastern A SEQUENCE-STRATIGRAPHIC et al., 2010) biostratigraphy. sector are older than rocks that occupy the same FRAMEWORK FOR THE DEPOSITION stratigraphic position in the east-central and OF THE LAFRAMBOISE MEMBER Deposition of the Oncolite Platform Bed western sectors. We infer geographic variations in the amount of erosion that occurred during the Using lithostratigraphy and biostratigraphy, The oncolite platform bed (Petryk, 1981) is hiatus, with less of the underlying Lousy Cove Desrochers et al. (2010) presented a sequence- a marker bed deÀ ning the base of the Lafram- Member preserved in the far east. Precise strati- stratigraphic model for the entire Ellis Bay boise Member (Long and Copper, 1987a). It graphic correlation of the sub-oncolitic Lousy Formation and identiÀ ed À ve transgressive- was identiÀ ed in all measured sections of the Cove Member across the island is critical to the regressive (TR) cycles, with the Laframboise Laframboise Member. We À nd signiÀ cant step- correct placement of biostratigraphically use- Member identiÀ ed as the À fth cycle (TR-5). 13 wise discontinuities in the I Ccarb measure- ful fossils and the interpretation of À rst and last Desrochers et al. (2010) showed that sediments across the base of the oncolite at all appearances of taxa during the end-Ordovician mentation during TR-1 through TR-4 was sections examined in the east-central and east- extinctions. diach ronous across the basin, challenging the ern sectors, but not in the west (Table 2). This east-west correlations proposed by Long and is chemostratigraphic evidence for a signiÀ cant Ellis Bay–Becscie Contact and the Copper (1987a). For example, the time line hiatus or erosional truncation at the base of the Ordovician-Silurian Boundary represented by the TR-1–TR-2 boundary oc- oncolite platform bed in the more nearshore curs in the Grindstone Member in the western sections. The base of the oncolite bed À lls in an The Ellis Bay–Becscie contact has been sug- sector, but it occurs in the Lousy Cove Member eroded and pitted surface with up to 10 cm of gested to represent the Ordovician-Silurian in the eastern sector. Diachroneity of facies is local relief, providing sedimentological evidence boundary (Cocks and Copper, 1981; Nowlan, also supported by chitino zoan biostratigraphy for discontinuous accumulation in all measured 1982; Long and Copper, 1987a; SouÀ ane and (Achab et al., 2010); the boundary between the 13 sections, although the I Ccarb continuity in the Achab, 2000; Copper, 2006; Melchin, 2008), H. crickmayi and B. gamachiana biozones oc- western sector suggests that not much time is but the Ordovician-Silurian boundary has also curs in the Grindstone Member in the west and missing there. We can use the absolute values been placed within the lower several meters of in the Lousy Cove Member in the east. 13 of the I Ccarb measurements of the base of the the Becscie Formation (McCracken and Barnes, The high-resolution chemostratigraphic data oncolite bed to track the timing of the deposi- 1981; Petryk, 1981; Copper, 1999; Desrochers presented here provide a means of intrabasinal tion of this unique facies. Using the ascending et al., 2010). As described already, in the west, correlation that exceeds the resolution available limb of the positive carbon isotope excursion as the basal Becscie Formation records a quick re- through lithostratigraphy and biostratigraphy. In 13 a time line (Brenchley et al., 2003), it appears turn in I Ccarb to its pre-excursion 0‰ baseline this section, we use the ascending and descend- 13 that the oncolite bed was deposited earlier in following the Hirnantian positive excursion. The ing limbs of the positive I Ccarb excursion as a 13 13 the western sector (where average I Ccarb of I Ccarb values in the lower Becscie of the east- chronometer to construct a chronology of indi- the oncolite is 2.0‰) than in the east-central central and eastern sectors stays elevated at +2‰ vidual rock units for all the measured sections 13 sector (where average I Ccarb of the oncolite is for at least 10 m above the Laframboise Member. of the Laframboise Member. We seek to use the

Geological Society of America Bulletin, July/August 2011 1657 Jones et al. D806

100 D805 Fox Point Member Point Fox Becscie Formation 80

13 D804/901 Figure 13. Composite I Ccarb 60 data from multiple measured sections of outcrops from English Head to Laframboise Point and Cape Henri, western Anticosti

Island. Data are from sections Height in section (m)

D821, D801–D806, and 901. Member Cove Lousy LFB See Figure 1 for section locations and Figure 8 for legend. 40 D803 Filled data points are from Ellis Bay Formation; open data points are from Becscie Forma- Formation Ellis Bay tion. Member boundaries below the Laframboise Member have Prinsta not been formalized on the D802 west coast and are used here 20 informally. LFB—Laframboise Member, VPDB—Vienna Peedee belemnite. Velleda D801

0 Grind- stone

(~150 m gap)

10 D821 Homards Member Homards Vauréal Formation Vauréal Height in section (m) 0 –5–4 –3 –2 –1 0241 3 5 18 13 δ Ocarb, δ Ccarb (‰, VPDB)

1658 Geological Society of America Bulletin, July/August 2011 Anticosti Island chemostratigraphy and sequence stratigraphy

δ13 TABLE 2. SUMMARY OF Ccarb DATA ACROSS THE LOUSY COVE–LAFRAMBOISE form bed itself is a transgressive lag deposit. CONTACT, INDICATING A DISCONFORMITY AT THE CONTACT Desrochers et al. (2010) also viewed the onco- Uppermost Lousy Cove Member Base of Laframboise Member oncolite Difference Section (‰, VPDB) (‰, VPDB) (‰, VPDB) lite as a transgressive lag. 901/D804 1.8 2.2 0.4 The Laframboise Member above the trans- 920 1.9 2.9 1.0 gressive surface and, in the east-central sector, D816 1.9 2.9 1.0 the resistant sandstones of the lowermost Fox 909A 2.0 2.9 1.0 Point Member together comprise the transgres- 909B 2.1 2.9 0.8 sive systems tract (TST). The isotope record 912A 2.2 2.9 0.7 at the sandstone-grainstone contact within the 914A 0.5 2.4 1.9 lower Fox Point Member shows a sharp dis- 914B 0.5 3.8 3.3 continuity in I13C in the east-central sec- Note: VPDB—Vienna Peedee belemnite. carb tor, indicating a brief period of nondeposition in these sections, perhaps due to carbonate sediment starvation as sea level rose rapidly. improved temporal resolution available through identiÀ ed in Figure 12 and yields a chronostrati- Depo si tion of the Fox Point sandstones ended chemostratigraphy to develop a more detailed graphic diagram with high temporal resolution diachronously within the east-central sector, oc- model for the deposition of the Lousy Cove, (Fig. 14). The vertical scale in Figure 14 is con- curring À rst in the westernmost section (Salmon Laframboise, and Fox Point Members and in structed linearly for clarity, but we note that the River upstream) and last in the easternmost sec- 13 doing so reÀ ne the chronology of glacioeustatic rate of change of I Ccarb was probably variable tion (Macaire Creek), supporting a sediment sea-level change, carbon cycle perturbation, and with time. The chronostratigraphic diagram pro- starvation hypothesis for the hiatus. In the far 13 extinction. vides a basis for developing a sequence-strati- western sector, at Laframboise Point, I Ccarb Disconformities in the succession are observ- graphic model for deposition of the Lousy Cove, stratigraphy indicates that a highly condensed able at the outcrop scale, and we can conÀ rm Laframboise, and Fox Point Members. Because sedimentary package was deposited with the their locations by identifying discontinuities in the Anticosti Basin carbonate ramp dipped to the bioherms. In this deeper, more distal location, 13 I Ccarb across the unconformable surfaces. The west, changes in sea level correspond to migra- slow in situ carbonate production in relatively surface separating the top of the Lousy Cove tion of facies belts in an east-west direction. calm waters may have facilitated the accumula- mudstones from the sandy carbonate grain- The Lousy Cover Member was deposited tion of this fairly complete condensed section. stones (in the west and east-central sectors) or under relatively deep waters in all parts of the This interpretation of the western sector con- the oncolite platform bed (in the east sector) island during the highstand systems tract (HST), tact between the Laframboise and Fox Point is one disconformity. In addition, the base of and we identify the disconformity at the top of the Members contrasts with that of Desrochers the oncolite platform bed in the Laframboise Lousy Cove Member mudstones as the sequence et al. (2010), who favored an exposure surface Member and the base of the Fox Point Member boundary (SB) of a depositional sequence. This and meters of erosion. Because we observe an were previously identiÀ ed as disconformities is the surface representing the start of sea-level interÀ ngering relationship between the top of (Brenchley et al., 2003; Bergström et al., 2006; fall, above which the falling stage systems tract the bioherms and the oldest Becscie grainstones Desrochers et al., 2010). However, the nature (FSST) and/or lowstand systems tract (LST) de- (Fig. 4F), we conclude that sedimentation was of these disconformities is variable across the veloped. During sea-level fall, increased erosion continuous, though very slow, in this location. 13 island. The I Ccarb data show a relatively small and siliciclastic input can generate a disconform- Other than the slow growth of the bioherms, discontinuity across the sub-oncolitic surface in able surface topped by reworked sediment; this the west was sediment starved until the pro- the west, a moderate increase in the east-central is our interpretation of the cross-bedded sandy grading Fox Point Member provided additional region, and a large step function in the east. A carbonate grainstone package that separates the sediment, as discussed later herein. This interval nearly opposite geographic pattern is identiÀ ed Lousy Cove mudstones from the oncolite plat- of sediment starvation may have produced the 13 at the base of the Becscie Formation; the I Ccarb form bed in the western and east-central sectors hardground surface capping the Laframboise discontinuity across the Ellis Bay–Becscie con- (Fig. 14). The FSST/LST is completely absent Member at Laframboise Point that Desrochers tact is nonexistent or very small in the east and in the eastern sector, probably due to subaerial et al. (2010) cited as evidence for subaerial ex- grows in magnitude in the east-central sections. exposure and erosion. posure. Sediment starvation and hardground At the westernmost section, high-resolution The chronostratigraphic diagram shows development at the distal western section are sampling and chemostratigraphy demonstrate that the oncolite platform bed was deposited consistent with maximum Á ooding conditions that there is not a large hiatus at the formation diachronously, with deposition beginning in the during this interval. boundary; the surface separating the Lafram- west and proceeding to the east. This tempo- Carbon isotope stratigraphy demonstrates boise Member from the Fox Point Member ral pattern is likely the result of rising sea-level that the base of the skeletal grainstones of the does not represent a profound unconformity, transgressing up the carbonate ramp from the lower Fox Point Member is diachronous across but rather a submarine hardground related to more distal western end to the more proximal Anticosti Island, with deposition beginning in sediment starvation. eastern end. We identify the base of the oldest the east and progressing westward over time 13 Because the positive I Ccarb excursion is oncolite (in the western sector) as the trans- (Fig. 14). We interpret the base of the Becscie a time-varying signal, we use it as a proxy gressive surface (TS) that marks the begin- Formation in the eastern sector as the maximum for time. Scaling the lithostratigraphy against ning of the transgressive systems tract (TST). Á ooding surface (MFS), the point at which sedi- 13 I Ccarb distributes individual packages of rock in The hiatus below the oncolite platform bed in ment accumulation began to outpace increase time (vertical dimension of Fig. 14). Plotting the the east-central and eastern sectors is likely the in sea-level rise. Sediments therefore prograded scaled measured sections as a function of east- result of transgressive ravinement as the shore- out across the ramp, À lling in the proximal basin west location on the island reinforces the trends line moved up the ramp, and the oncolite plat- À rst and the more distal parts of the basin later.

Geological Society of America Bulletin, July/August 2011 1659 Jones et al.

Distal Proximal West Central East-central East

0.0 0.5 Younger

1.0 HST 1.5 2.0 Hard- Missing Missing 2.5 ground strata strata? 3.0 MFS

excursion excursion 3.5 TST

carb 4.0 C

13 3.5 δ 3.0 2.5

2.0 FSST/LST TS

Hirnantian Eroded/bypass 1.5 1.0 0.5 SB

0.0 HST 0.0 Older Laframboise Point Salmon River Salmon River Macaire Creek Lousy Cove Low High (D804, 901) upstream downstream (912, D815) (914) Sea level (D816, D817, 911) (D814, 909A/B)

Becscie Formation Ellis Bay Formation Fox Point Laframboise Lousy Cove Member Member Member Oncolite platform

LEGEND bed Bioherm

13 Figure 14. Chronostratigraphic diagram of upper Lousy Cove, Laframboise, and Lower Fox Point Members based on I Ccarb stratigraphy 13 at À ve sections in an east-west transect of Anticosti Island. Vertical axis is I Ccarb (‰) during the Hirnantian positive isotope excursion. Although the vertical axis has been plotted on a linear scale, subdivisions in the isotope axis probably do not represent equal intervals of time. There are no adequate outcrop exposures for data in the central portion of the island. SB—sequence boundary, TS—transgressive surface, MFS—maximum Á ooding surface, HST—highstand systems tract, FSST—falling stage systems tract, LST—lowstand systems tract, TST—transgressive systems tract.

As such, the bulk of the Fox Point Member rep- IMPLICATIONS FOR THE PLACEMENT ences about the cause of the isotope excursion. resents the highstand systems tract (HST) of the OF THE ORDOVICIAN-SILURIAN Furthermore, our sequence-stratigraphic model depositional sequence. BOUNDARY AND INTERPRETATION implies that similar facies in different sectors of The sequence-stratigraphic model outlined OF LINKAGES AMONG GLACIATION, the island are not chronostratigraphic equiva- here, based on integrated lithostratigraphy and ISOTOPE EXCURSIONS, AND lents, especially in the strata adjacent to and carbon isotope chemostratigraphy, implies that MASS EXTINCTION within the Laframboise Member. This observa- maximum glacioeustatic sea-level fall occurred tion has bearing on the temporal distribution of at the top of the Lousy Cove mudstones, and The integrated sequence-stratigraphic and À rst and last appearances of fossils in the basin, that sea level was rising during the deposition chemostratigraphic framework presented here and therefore may complicate interpretations of the Laframboise and Fox Point Members. provides new constraints on end-Ordovician of the end-Ordovician mass extinction on Anti- The consequences of this interpretation for end- history. The relative timing of the glacioeustatic costi Island. The base of the Silurian at Dob’s 13 Ordovician Earth history are developed in the sea-level changes and the initiation and termina- Linn is isotopically characterized by I Corg val- next section. tion of the I13C excursion allow us to make infer- ues at pre-excursion baseline. The excursion is

1660 Geological Society of America Bulletin, July/August 2011 Anticosti Island chemostratigraphy and sequence stratigraphy

13 over by the start of the Silurian. Our data there- is organic carbon) and leads to elevated I Ccarb forcing crossed the pCO2 threshold needed to fore imply that, from the standpoint of isotope values. Second, increased Forg results in a draw- deglaciate Gondwana. As mentioned already, stratigraphy, the Ordovician-Silurian boundary down of pCO2 due to increased carbon À xa- there is no support for a substantive reproduci- 13 is at the Ellis Bay–Becscie contact only in the tion. This drop in pCO2 is, in turn, the proposed ble change in ) C in our Anticosti data and, as westernmost sections at Laframboise Point. driver for Hirnantian glaciation in this scenario such, no geochemical evidence for an increase 13 13 Farther to the east, I Ccarb and I Corg values are (Marshall and Middleton, 1990; Brenchley in pCO2 spanning the Hirnantian glaciation and elevated above baseline levels in the lower ex- et al., 1994). The productivity hypothesis does I13C excursion. posures of the Fox Point Member of the Becscie not address the ultimate cause of the proposed The “weathering” hypothesis (Kump et al.,

Formation, a consequence of the diachronous increase in Forg. Based on the relationship be- 1999; Melchin and Holmden, 2006) also pro- deposition during the basinward progradation of tween pCO2 and ice volume in the Pleistocene, vides a separate explanation for the cause of the 13 the Fox Point Member. it is possible that a change in pCO2 was driven I Ccarb excursion. In this scenario, the positive 13 We infer that the forced regression during by changing ocean circulation associated with I Ccarb excursion is due to the enhanced weath- the falling stage systems tract was a response to the onset or termination of glacial conditions ering of freshly exposed carbonate rocks dur- glacioeustatic sea-level drawdown due to devel- (i.e., pCO2 driven by glacial changes rather ing glacioeustatic sea-level fall. Isotope mass opment of maximum glacial condition on Gond- than vice versa). Given that we still lack a clear balance models (Kump et al., 1999; Kump 13 wana, and the sea-level rise associated with the under standing of the causal relationship be- and Arthur, 1999) show that I Ccarb rises if the transgressive systems tract was caused by melt- tween pCO2 and ice volume during the Pleisto- freshly exposed material has a higher ratio of ing back of the Gondwana glaciers (Fig. 14). As cene (Sigman and Boyle, 2000), it is likely that carbonate carbon to organic carbon than what established by Desrochers et al. (2010), Anti- any relationship between pCO2 and glaciation is was weathering previously. As such, Hirnantian costi Island records À ve transgressive-regressive complicated and particularly difÀ cult to recon- glacioeustatic sea-level fall would have resulted 13 sequences throughout the Ellis Bay Formation. struct without a robust paleo-pCO2 proxy. We in an increase in the I C composition of the Desrochers et al. (2010) interpreted the À rst four note that )13C has been proposed as a proxy for riverine input to the oceans. This explanation 13 TR cycles as ~400 k.y. eccentricity-controlled reconstructing paleo-pCO2 (Arthur et al., 1985; for the I Ccarb excursion does not require any glacial cycles and suggested that the À fth TR Popp et al., 1989; Laws et al., 1995). However, change in forg, the organic carbon burial fraction. cycle , corresponding to the Laframboise Mem- no substantive reproduci ble change in )13C is An increase in riverine I13C is consistent with 13 13 ber, was deposited during a brief inter glacial found in our data, nor in the majority of Hirnan- the observed I Ccarb and I Corg data presented period. Our model offers an alternate interpre- tian chemostratigraphic studies (e.g., LaPorte here and reported elsewhere. tation for the most severe glacioeustatic sea- et al., 2009; but see Young et al. [2010] for an We are left with two contrasting explanations level change in the Ellis Bay Formation (TR-5 alternate view). Further, )13C is believed to re- for the origin of the Hirnantian I13C excursion, of Desrochers et al., 2010), associated with the Á ect many other parameters and, as such, does one due to increased organic carbon burial (Mar-

Laframboise Member. We suggest that what not constitute a robust pCO2 indicator. This leads shall and Middleton, 1990; Brenchley et al., makes the Laframboise Member lithologically us to conclude that a complex set of factors was 1994) and the other due to increased I13C of riv- distinctive is the unique environment of its controlling )13C on a spatial scale much smaller erine input (Kump et al., 1999). Both of these 13 depo si tion: the reÁ ooding of the Anticosti Basin than the size of the Anticosti Basin. We do not scenarios result in parallel increases in I Ccarb 13 13 during terminal Ordovician deglaciation. make any global scale inferences from the ) C and I Corg, and our paired geochemical data on Published studies of carbon and oxygen iso- data and caution against such speculation in their own cannot discriminate between them. topes from carbonate rocks on Anticosti Island studies that do not reproduce )13C signals from It is possible, however, to use the relationship (Orth et al., 1986; Long, 1993; Brenchley et al., multiple locations within a basin. In summary, between the carbon isotopic excursion and our 18 1994) have identiÀ ed enriched I Ocarb and there is no À rm geochemical evidence based on reconstructed sea-level curve to further con- 13 13 I Ccarb in the Laframboise Member of the Ellis Anticosti ) C for a decrease in pCO2 associ- strain possible mechanisms for generating the Bay Formation. The oxygen data have been in- ated with the Hirnantian glaciation and I13C Hirnantian I13C excursion. The identiÀ cation terpreted to represent a combination of cooling excursion. On the other hand, an increase in forg of the beginning and end of the maximal pulse 13 ocean temperatures and continental ice-sheet remains consistent with the observed I Ccarb of end-Ordovician glaciation in the Anticosti 13 growth associated with terminal Ordovician and I Corg data presented here and reported stratigraphic record allows us to determine the glaciation on Gondwana (Marshall and Middle- elsewhere. relative timing of the carbon isotope excursion ton, 1990; Long, 1993; Brenchley and Marshall, The “weathering” hypothesis (Kump et al., and the culmination of the ice age. On Anticosti, 1999; Brenchley et al., 2003). The synchrony 1999; Melchin and Holmden, 2006) was origi- the start of the ascending limb of the I13C ex- 18 13 of the I Ocarb and I Ccarb excursions has led to nally formulated to explain the aforementioned cursion coincides with the sequence boundary 13 13 hypotheses linking climate change and positive positive ) C coincident with the I Ccarb excur- at the top of the Lousy Cove Member (Fig. 14). carbon isotope excursions. Foremost among the sion observed in Monitor Range, Nevada. In this The “productivity hypothesis” (Marshall and competing models are the “productivity” and scenario, atmospheric pCO2 is thought to have Middleton, 1990; Brenchley et al., 1994; Mar- “weathering” hypotheses, and we next discuss decreased leading into the glaciations, and the shall et al., 1997) predicts that increased carbon 13 these hypotheses in the context of our observa- subsequent ) C excursion is interpreted as evi- burial lowered pCO2 until a climatic threshold tions of Anticosti stratigraphy. dence of increasing pCO2 due to reduced silicate was crossed that allowed polar glaciers to de- The “productivity” hypothesis (Marshall weathering during the time that continental ice velop. At the resolution provided by sequence and Middleton, 1990; Brenchley et al., 1994) was covering silicate rock terranes during glaci- stratigraphy and chemostratigraphy, we see no posits an increase in Forg (the magnitude of the ation until greenhouse forcing crossed a thresh- evidence for an increase in organic carbon burial total organic carbon burial Á ux). This has two old needed to deglaciate Gondwana. Kump (as indicated by a positive I13C excursion) be- direct effects (Kump and Arthur, 1999). First, it et al. (1999) posited that pCO2 increased over fore the initiation of maximum continental ice- raises forg (the fraction of total carbon buried that the course of the glaciation until greenhouse sheet growth and glacioeustatic sea-level fall.

Geological Society of America Bulletin, July/August 2011 1661 Jones et al.

Our interpretation of the paired sequence- Integrated lithostratigraphy and isotope versity: Paleobiology, v. 30, p. 522–542, doi: 10.1666/ 0094-8373(2004)030<0522:OEAMDO>2.0.CO;2. stratigraphic and carbon isotope data also im- chemo stratigraphy form the foundation of a se- Banner, J.L., and Hanson, G., 1990, Calculation of simul- 13 plies that the I C excursion peaked during quence-stratigraphic model for deposition of the taneous isotopic and trace element variations during maximum glacial conditions and that (deglacial) Laframboise Member. Glacioeustatic sea-level water-rock interaction with applications to carbonate 13 diagenesis: Geochimica et Cosmochimica Acta, v. 54, sea-level rise had begun before the I C excur- fall caused a forced regression that deposited the p. 3123–3137, doi: 10.1016/0016-7037(90)90128-8. sion ended. The maximum Á ooding surface oc- cross-stratiÀ ed sand beds beneath the oncolite Barnes, C.R., 1988, Stratigraphy and palaeontology of the curs when I13C is ~3‰, on the descending platform bed. ReÁ ooding of the ramp during de- Ordovician-Silurian boundary interval, Anticosti Is- carb land, Québec, Canada: Bulletin of the British Museum limb of the excursion (Fig. 14). This relative glaciation led to deposition of the transgressive of Natural History (Geology), v. 43, p. 195–219. timing is consistent with the two aforemen- oncolite platform bed (transgressive lag of the Bergström, S.M., Saltzman, M.R., and Schmitz, B., 2006, 13 13 First record of the Hirnantian (Upper Ordovician) I C tioned mechanisms for increased I C during the transgressive systems tract) and the remainder excursion in the North American Midcontinent and its Hirnantian: elevated forg (Brenchley et al., 1994) of the Laframboise Member and lower Becscie regional implications: Geological Magazine, v. 143, and increased I13C of riverine input (Kump et al., sandstones. As sea-level rise slowed, skeletal p. 657–678, doi: 10.1017/S0016756806002469. Bergström, S.M., Chen, X., Gutierrez-Marco, J.C., and 1999), allowing for the time lag corresponding to grainstones of the Fox Point Member (highstand Dronov, A., 2009, The new chronostratigraphic clas- a change in carbon cycling to propagate through systems tract) prograded basinward, from east siÀ cation of the Ordovician System and its rela- the ocean system (Kump and Arthur, 1999). The to west. tions to major regional series and stages and to I13C chemostratigraphy: Lethaia, v. 42, p. 97–107, doi: constraints on the timing of latest Ordovician The observed relative timing between sea- 10.1111/j.1502-3931.2008.00136.x. glacioeustatic sea-level changes provided here level fall and the beginning of the I13C ex- Bergström, S.M., Young, S., and Schmitz, B., 2010, Katian (Upper Ordovician) I13C chemostratigraphy and se- may provide new insights for future modeling cursions is consistent with the “weathering” quence stratigraphy in the United States and Balto- efforts focused on the origin of the Hirnantian hypothesis advanced to explain the origin of scandia: A regional comparison: Palaeogeography, 13 Palaeoclimatology, Palaeoecology, v. 296, p. 217–234, positive carbon isotope excursion. the Hirnantian positive I Ccarb excursion. The doi: 10.1016/j.palaeo.2010.02.035. excursion persisted through deglacial sea-level Berner, R.A., 2006, GEOCARBSULF: A combined model

CONCLUSIONS rise, indicating that the response time for the for Phanerozoic atmospheric O2 and CO2: Geochi- Hirnantian carbon cycle perturbation is discern- mica et Cosmochimica Acta, v. 70, p. 5653–5664, doi: 10.1016/j.gca.2005.11.032. Our stratigraphic and isotopic studies of ible in the Anticosti stratigraphic record. Bordet, E., Malo, M., and Kirkwood, D., 2010, A struc- the carbonate rocks spanning the Ordovician- tural study of western Anticosti Island, St. Lawrence ACKNOWLEDGMENTS platform, Québec: A fracture analysis that integrates Silurian boundary across Anticosti Island sug- surface and subsurface structural data: Bulletin of gest the following. Canadian Petroleum Geology, v. 58, p. 36–55, doi: We are grateful to Bob Criss and Liz Hasenmueller 10.2113/gscpgbull.58.1.36. The globally documented positive carbon 13 18 for help with the I Ccarb and I Ocarb measurements Brenchley, P.J., and Marshall, J.D., 1999, Relative timing isotope excursion during the Hirnantian Stage is at Washington University. Greg Eischeid supervised and critical events during the Late Ordovician mass recorded in both carbonate and organic carbon the mass spectrometry laboratory at Harvard Univer- extinction—New data from Oslo: Acta Universitatis at the Ellis Bay–Becscie contact. The magnitude sity. We beneÀ ted from Jon Husson’s assistance in Carolinae-Geologica, v. 43, p. 187–190. Brenchley, P.J., Marshall, J.D., Carden, G.A.F., Robertson, of the excursion is ~3‰–4‰ in both phases. the À eld and thank Megan Rohrsson, Tim Raub, and Priya Nayak for valuable discussions at the outcrops. D.B.R., Long, D.G.F., Meidla, T., Hints, L., and Ander- Detailed chemostratigraphy of the upper Vau- We thank the owners of the Salmon River Lodge for son, T.F., 1994, Bathymetric and isotopic evidence for a short-lived Late Ordovician glaciation in a greenhouse réal and the Ellis Bay Formations reveals a small access to outcrops on the Salmon River. Dave Boulet period: Geology, v. 22, p. 295–298, doi: 10.1130/ 13 positive I Ccarb excursion below the Lafram- and SEPAQ Anticosti granted permission to work in 0091-7613(1994)022<0295:BAIEFA>2.3.CO;2. boise Member, with tens of meters of carbonate Anticosti National Park. We thank Paul Copper for Brenchley, P.J., Marshall, J., and Underwood, C., 2001, Do suggesting outcrops to visit and for generously shar- all mass extinctions represent an ecological crisis? Evi- strata at 0‰ in between. This suggests that the ing his extensive knowledge of Anticosti geology. dence from the Late Ordovician: Geological Journal, Laframboise–lower Fox Point Members con- We thank Associate Editor Maya Elrick, Michael v. 36, p. 329–340, doi: 10.1002/gj.880. tain a complete record of the Hirnantian Stage, Joachimski, André Desrochers, and an anonymous Brenchley, P.J., Carden, G.A.F., Hints, L., Kaljo, D., Mar- shall, J.D., Martma, T., Meidla, T., and Nõlvak, J., 13 referee for their helpful and thorough reviews of the and the lower I Ccarb peaks may correspond to 2003, High-resolution stable isotope stratigraphy of a Katian excursion documented in Baltica and manuscript. Paul Hoffman provided inspiration and Upper Ordovician sequences: Constraints on the tim- guidance in the early phases of the project. The project ing of bioevents and environmental changes associated Laurentia. was supported by a grant from the Agouron Institute with mass extinction and glaciation: Geological Society 13 Using the ascending limb of the I Ccarb ex- to Fike and Fischer and a grant from the Henry Breck of America Bulletin, v. 115, p. 89–104, doi: 10.1130/ cursion as a time-varying signal, we interpret Fund through the Harvard University Center for the 0016-7606(2003)115<0089:HRSISO>2.0.CO;2. Environment to Jones. Caputo, M.V., and Crowell, J.C., 1985, Migration of gla- the base of the oncolite platform bed (Lafram- cial centers across Gondwana during Paleozoic boise Member, Ellis Bay Formation) to be Era: Geological Society of America Bulletin, v. 96, REFERENCES CITED time transgressive, with deposition progressing p. 1020–1036, doi: 10.1130/0016-7606(1985)96<1020 :MOGCAG>2.0.CO;2. across the basin from west to east. Achab, A., Asselin, E., Desrochers, A., Riva, J.F., and Farley, Chen, X., Rong, J., Fan, J., Zhan, R., Mitchell, C.E., Harper, The smooth nature of the descending limb of C., 2011, Chitinozoan contribution to the development D.A.T., Melchin, M.J., Peng, P., Finney, S.C., and the 13C curve at Laframboise Point suggests of a new Upper Ordovician stratigraphic framework for Wang, X., 2006, The global boundary stratotype sec- I carb Anticosti Island: Geological Society of America Bulle- tion and point (GSSP) for the base of the Hirnantian that there is not a major hiatus across the Ordo- tin, v. 123, p. 186–205, doi: 10.1130/B30131.1. Stage (the uppermost of the Ordovician System): Epi- vician-Silurian boundary in this section. Ainsaar, L., Kaljo, D., Martma, T., Meidla, T., Männik, P., sodes, v. 29, p. 183–196. Nõlvak, J., and Tinn, O., 2010, Middle and Upper Cocks, L.R.M., and Copper, P., 1981, The Ordovician- The skeletal grainstones at the base of the Ordovician carbon isotope chemostratigraphy in Silurian boundary at the eastern end of Anticosti Fox Point Member (Becscie Formation) are Balto scandia: A correlation standard and clues to envi- Island: Canadian Journal of Earth Sciences, v. 18, diachronous across the island, with deposition ron mental history: Palaeogeography, Palaeoclimatol- p. 1029–1034. ogy, Palaeoecology, v. 294, p. 189–201, doi: 10.1016/ Copper, P., 1999, Brachiopods during and after the Late progressing from east to west. The Ordovician- j.palaeo.2010.01.003. Ordovician mass extinctions, Anticosti Island, E Silurian boundary likely occurs tens of meters Arthur, M.A., Dean, W.E., and Claypool, G.E., 1985, Anom- Canada: Acta Universitatis Carolinae-Geologica, v. 43, 13 into the lower Fox Point Member in the east, alous C enrichment in modern marine organic carbon: p. 207–209. Nature, v. 315, p. 216–218, doi: 10.1038/315216a0. Copper, P., 2001, Reefs during the multiple crises towards and 1–2 m above the Laframboise–Fox Point Bambach, R.K., Knoll, A.H., and Wang, S.C., 2004, Origi- the Ordovician-Silurian boundary: Anticosti Island, contact in the west. nation, extinction, and mass depletions of marine di- eastern Canada, and worldwide: Canadian Journal

1662 Geological Society of America Bulletin, July/August 2011 Anticosti Island chemostratigraphy and sequence stratigraphy

of Earth Sciences, v. 38, p. 153–171, doi: 10.1139/ and mineral study, diagenesis and petroleum potential: land basin, Québec, Canada: Canadian Journal of Earth cjes-38-2-153. Sainte-Foy, Quebec, Institut National de la Recherche Sciences, v. 44, p. 413–431, doi: 10.1139/E06-099. Copper, P., 2006, Faunal change across the Ordovician- ScientiÀ que, 7, 27 p. Long, D.G.F., and Copper, P., 1987a, Stratigraphy of the Silurian mass extinction boundary on Anticosti Island, Jin, J., and Copper, P., 2008, Response of brachiopod com- Upper Ordovician upper Vauréal and Ellis Bay For- E Canada: Geological Society of America Abstracts munities to environmental change during the Late mations, eastern Anticosti Island, Québec: Canadian with Programs, v. 38, no. 7, p. 403. Ordovician mass extinction interval, Anticosti Island, Journal of Earth Sciences, v. 24, p. 1807–1820, doi: Copper, P., and Long, D.G.F., 1989, Stratigraphic revisions eastern Canada: Fossils and Strata, v. 54, p. 41–51. 10.1139/e87-172. for a key Ordovician/Silurian boundary section, Anti- Kaljo, D., Hints, L., Martma, T., and Nõlvak, J., 2001, Long, D.G.F., and Copper, P., 1987b, Late Ordovician sand- costi Island, Canada: Newsletters on Stratigraphy, Carbon isotope stratigraphy in the latest Ordovician wave complexes on Anticosti Island, Québec: A marine v. 21, p. 59–73. of Estonia: Chemical Geology, v. 175, p. 49–59, doi: tidal embayment?: Canadian Journal of Earth Sciences, Delabroye, A., and Vecoli, M., 2010, The end-Ordovician 10.1016/S0009-2541(00)00363-6. v. 24, p. 1821–1832, doi: 10.1139/e87-173. glaciation and the Hirnantian Stage: A global review Kaljo, D., Hints, L., Martma, T., Nõlvak, J., and Oraspld, Marshall, J.D., and Middleton, P.D., 1990, Changes in and questions about Late Ordovician event stratig- A., 2004, Late Ordovician carbon isotope trend in marine isotopic composition and the late Ordovician raphy: Earth-Science Reviews, v. 98, p. 269–282, Estonia, its signiÀ cance in stratigraphy and environ- glaciation: Journal of the Geological Society of Lon- doi: 10.1016/j.earscirev.2009.10.010. mental analysis: Palaeogeography, Palaeoclimatol- don, v. 147, p. 1–4, doi: 10.1144/gsjgs.147.1.0001. Desrochers, A., and Gauthier, E.L., 2009, Carte Géologique ogy, Palaeoecology, v. 210, p. 165–185, doi: 10.1016/ Marshall, J.D., Brenchley, P.J., Mason, P.R.D., Wolff, G.A., Synthèse de l’Île d’Anticosti. Annotated Map: Québec, j.palaeo.2004.02.044. Astini, R.A., Hints, L., and Meidla, T., 1997, Global Ministère des Ressources Naturelles et de la Faune, Kaljo, D., Hints, L., Maennik, P., and Nõlvak, J., 2008, The carbon isotopic events associated with mass extinction scale 1:250,000. succession of Hirnantian events based on data from and glaciation in the Late Ordovician: Palaeogeog- Desrochers, A., Farley, C., Achab, A., Asselin, E., and Riva, Baltica: Brachiopods, chitinozoans, conodonts, and raphy, Palaeoclimatology, Palaeoecology, v. 132, J.F., 2010, A far-À eld record of the end Ordovician carbon isotopes: Estonian Journal of Earth Science, p. 195–210, doi: 10.1016/S0031-0182(97)00063-1. glaciation: The Ellis Bay Formation, Anticosti Island, v. 57, p. 197–218, doi: 10.3176/earth.2008.4.01. McCracken, A.D., and Barnes, C.R., 1981, Conodont bio- Eastern Canada: Palaeogeography, Palaeoclimatol- Kamo, S.L., and Gower, C., 1994, Note: U-Pb baddeleyite stratigraphy and paleoecology of the Ellis Bay Forma- ogy, Palaeoecology, v. 296, p. 248–263, doi: 10.1016/ dating clariÀ es age of characteristic paleomagnetic tion, Anticosti Island, Québec, with special reference j.palaeo.2010.02.017. rema nence of Long Range dykes, southeastern Labra- to Late Ordovician–Early Silurian chronostratigraphy Díaz-Martínez, E., and Grahn, Y., 2007, Early Silurian gla- dor: Atlantic Geology, v. 30, p. 259–262. and the systematic boundary: Geological Survey of ciation along the western margin of Gondwana (Peru, Kump, L.R., and Arthur, M.A., 1999, Interpreting carbon- Canada Bulletin, v. 329, p. 51–111. Bolivia and northern Argentina): Palaeogeographic and isotope excursions: Carbonates and organic matter: Melchin, M.J., 2008, Restudy of some Ordovician-Silurian geodynamic setting: Palaeogeography, Palaeoclimatol- Chemical Geology, v. 161, p. 181–198, doi: 10.1016/ boundary graptolites from Anticosti Island, Canada, ogy, Palaeoecology, v. 245, p. 62–81, doi: 10.1016/ S0009-2541(99)00086-8. and their biostratigraphic signiÀ cance: Lethaia, v. 41, j.palaeo.2006.02.018. Kump, L.R., Arthur, M.A., Patzkowsky, M., Gibbs, M., p. 155–162, doi: 10.1111/j.1502-3931.2007.00045.x. Emiliani, C., 1966, Isotopic paleotemperatures: Science, Pinkus, D.S., and Sheehan, P.M., 1999, A weathering Melchin, M.J., and Holmden, C.E., 2006, Carbon isotope

v. 154, p. 851–857, doi: 10.1126/science.154.3751.851. hypothesis for glaciation at high atmospheric pCO2 chemostratigraphy in Arctic Canada: Sea-level forc- Fan, J., Peng, P., and Melchin, M.J., 2009, Carbon isotopes during the Late Ordovician: Palaeogeography, Palaeo- ing of carbonate platform weathering and implications and event stratigraphy near the Ordovician-Silurian climatology, Palaeoecology, v. 152, p. 173–187, doi: for Hirnantian global correlation: Palaeogeography, boundary, Yichang, South China: Palaeogeography, 10.1016/S0031-0182(99)00046-2. Palaeoclimatology, Palaeoecology, v. 234, p. 186–200, Palaeoclimatology, Palaeoecology, v. 276, p. 160–169, Lake, J.H., 1981, Sedimentology and paleoecology of Upper doi: 10.1016/j.palaeo.2005.10.009. doi: 10.1016/j.palaeo.2009.03.007. Ordovician mounds of Anticosti Island, Québec: Cana- Melchin, M.J., and Williams, S.H., 2000, A restudy of the Farley, C., and Desrochers, A., 2007, Sediment dynamics dian Journal of Earth Sciences, v. 18, p. 1562–1571. Akidograptine graptolites from Dob’s Linn and a pro- and stratigraphic architecture of a mixed carbonate- LaPorte, D.F., Holmden, C.E., Patterson, W.P., Loxton, posed redeÀ ned zonation of the Silurian stratotype: siliciclastic ramp: The Upper Ordovician (Hirnan- J.D., Melchin, M.J., Mitchell, C.E., Finney, S.C., and Palaeontology Down-Under, v. 61, p. 63. tian) Ellis Bay Formation, Anticosti Island, Québec, Sheets, H.D., 2009, Local and global perspectives Nowlan, G.S., 1982, Conodonts and the position of the Canada: Geological Society of America Abstracts with on carbon and nitrogen cycling during the Hirnan- Ordovician-Silurian boundary at the eastern end of Programs, v. 39, p. 419. tian glaciation: Palaeogeography, Palaeoclimatology, Anticosti Island, Québec, Canada: Canadian Journal Finney, S.C., Berry, W.B.N., Cooper, J.D., Ripperdan, R.L., Palaeoecology, v. 276, p. 182–195, doi: 10.1016/ of Earth Sciences, v. 19, p. 1332–1335. Sweet, W.C., Jacobson, S.R., SouÀ ane, A., Achab, A., j.palaeo.2009.03.009. Orth, C.J., Gilmore, J.S., Quintana, L.R., and Sheehan, P.M., and Noble, P.J., 1999, Late Ordovician mass extinction: Laws, E.A., Popp, B.N., Bidigare, R.R., Kennicutt, M.C., 1986, Terminal Ordovician extinction; geochemical A new perspective from stratigraphic sections in cen- and Macko, S.A., 1995, Dependence of phytoplank- analysis of the Ordovician/Silurian boundary, Anticosti tral Nevada: Geology, v. 27, p. 215–218, doi: 10.1130/ ton carbon isotopic composition on growth rate and Island, Québec: Geology, v. 14, p. 433–436, doi: 10.1130/

0091-7613(1999)027<0215:LOMEAN>2.3.CO;2. [CO2]aq: Theoretical considerations and experimental 0091-7613(1986)14<433:TOEGAO>2.0.CO;2. Ghienne, J.-F., 2003, Late Ordovician sedimentary environ- results: Geochimica et Cosmochimica Acta, v. 59, Patterson, W., and Walter, L., 1994, Depletion of 13C in

ments, glacial cycles, and post-glacial transgression p. 1131–1138, doi: 10.1016/0016-7037(95)00030-4. seawater 8CO2 on modern carbonate platforms: in the Taoudeni Basin, West Africa: Palaeogeography, Le Heron, D.P., and Dowdeswell, J., 2009, Calculating ice SigniÀ cance for the carbon isotopic record of car- Palaeoclimatology, Palaeoecology, v. 189, p. 117–145, volumes and ice Á ux to constrain the dimensions of a bonates: Geology, v. 22, p. 885–888, doi: 10.1130/ doi: 10.1016/S0031-0182(02)00635-1. 440 Ma North African ice sheet: Journal of the Geo- 0091-7613(1994)022<0885:DOCISC>2.3.CO;2. Grahn, Y., and Caputo, M.V., 1992, Early Silurian glacia- logical Society of London, v. 166, p. 277–281, doi: Petryk, A.A., 1981, Stratigraphy, sedimentology and paleo- tions in Brazil: Palaeogeography, Palaeoclimatology, 10.1144/0016-76492008-087. geography of the Upper Ordovician–Lower Silurian Palaeoecology, v. 99, p. 9–15, doi: 10.1016/0031-0182 Le Heron, D.P., and Howard, J., 2010, Evidence for Late of Anticosti Island, Québec, in Lesperance, P.J., ed., (92)90003-N. Ordovician glaciation of Al Kufrah Basin, Libya: Jour- Subcommission on Silurian Stratigraphy, Ordovician- Hambrey, M.J., 1985, The Late Ordovician–Early Silurian nal of African Earth Sciences, v. 58, p. 354–364, doi: Silurian Working Group. Volume II: Stratigraphy and glacial period: Palaeogeography, Palaeoclimatol- 10.1016/j.jafrearsci.2010.04.001. Paleontology: Anticosti-Gaspe, Québec, Universite de ogy, Palaeo ecology, v. 51, p. 273–289, doi: 10.1016/ Le Heron, D.P., Ghienne, J.-F., Houicha, M.E., Khoukhi, Y., Montreal, p. 11–39. 0031-0182(85)90089-6. and Rubino, J.L., 2007, Maximum extent of ice sheets Pinet, N., and Lavoie, D., 2007, The offshore part of the Hambrey, M.J., and Harland, W.B., eds., 1981, Earth’s Pre- in Morocco during the Late Ordovician glaciation: Anticosti Basin: A major gap in the understanding of Pleistocene Glacial Record: Cambridge, UK, Cam- Palaeo geography, Palaeoclimatology, Palaeoecology, early to middle Paleozoic basins of Eastern Canada in bridge University Press, 1024 p. v. 245, p. 200–226, doi: 10.1016/j.palaeo.2006.02.031. a promising hydrocarbon setting, in Convention Ex- Holland, C.H., 1985, Series and Stages of the Silurian Sys- Loi, A., Ghienne, J.F., Dabard, M.P., Paris, F., Botquelen, A., tended Abstracts: Calgary, Canadian Society of Petro- tem: Episodes, v. 8, p. 101–103. Christ, N., Elaouad-Debbaj, Z., Gorini, A., Vidal, M., leum and Geology, p. 336–339. Holmden, C.E., Creaser, R.A., Muehlenbachs, K., Leslie, Videt, B., and Destombes, J., 2010, The Late Ordovi- Popp, B., Takigiku, R., Hayes, J., Louda, J., and Baker, E., S.A., and Bergström, S.M., 1998, Isotopic evidence cian glacio-eustatic record from a high-latitude storm- 1989, The post-Paleozoic chronology and mechanism for geochemical decoupling between ancient epeiric dominated shelf succession: The Bou Ingarf section for 13C depletion in primary marine organic matter: seas and bordering oceans: Implications for secular (Anti-Atlas, southern Morocco): Palaeogeography, American Journal of Science, v. 289, p. 436–454, doi: curves: Geology, v. 26, p. 567–570, doi: 10.1130/ Palaeoclimatology, Palaeoecology, v. 296, p. 332–358, 10.2475/ajs.289.4.436. 0091-7613(1998)026<0567:IEFGDB>2.3.CO;2. doi: 10.1016/j.palaeo.2010.01.018. Ripperdan, R.L., Cooper, J.D., and Finney, S.C., 1998, High Immenhauser, A., Porta, G.D., Kenter, J.A.M., and Long, D.G.F., 1993, Oxygen and carbon isotopes and event resolution I13C and lithostratigraphic proÀ les from Bahamonde, J.R., 2003, An alternative model for stratigraphy near the Ordovician-Silurian boundary, Copen hagen Canyon, Nevada: Clues to the behavior positive shifts in shallow-marine carbonate I13C and Anticosti Island, Québec: Palaeogeography, Palaeo- of the ocean carbon cycle during the Late Ordovi- I18O: Sedimentology, v. 50, p. 953–959, doi: 10.1046/ climatology, Palaeoecology, v. 104, p. 49–59, doi: cian global crisis: Mineralogical Magazine, v. 62A, j.1365-3091.2003.00590.x. 10.1016/0031-0182(93)90119-4. p. 1279–1280, doi: 10.1180/minmag.1998.62A.3.03. Institut National de la Recherche Scientifique (INRS)- Long, D.G.F., 2007, Tempestite frequency curves: A key to Runkel, A.C., Mackey, T.J., Cowan, C.A., and Fox, D.L., Petrole, 1974, Arco-Anticosti No. 1: Sedimentologic, Late Ordovician and Early Silurian subsidence, sea- 2010, Tropical shoreline ice in the Late Cambrian: mineralogical, bio-stratigraphical, geochemical, organic level change, and orbital forcing in the Anticosti fore- Implications for Earth’s climate between the Cambrian

Geological Society of America Bulletin, July/August 2011 1663 Jones et al.

explosion and the Great Ordovician biodiversiÀ cation tigraphy of the basal Silurian stratotype (Dob’s Linn, ecology, v. 132, p. 147–158, doi: 10.1016/S0031-0182 event: GSA Today, v. 20, no. 11, p. 4–10, doi: 10.1130/ Scotland) and its global correlation: Journal of the (97)00046-1. GSATG84A.1. Geological Society of London, v. 154, p. 709–718, doi: Webby, B.D., Paris, F., Droser, M.L., and Percival, I.G., eds., Sami, T., and Desrochers, A., 1992, Episodic sedimenta- 10.1144/gsjgs.154.4.0709. 2004, The Great Ordovician BiodiversiÀ cation Event: tion on an Early Silurian, storm-dominated carbonate Urey, H.C., 1947, The thermodynamic properties of iso- New York, Columbia University Press, 496 p. ramp, Becscie and Merrimack Formations, Anticosti topic substances: Journal of the Chemical Society, Yan, D., Chen, D., Wang, Q., Wang, J., and Wang, Z., Island, Canada: Sedimentology, v. 39, p. 355–381, doi: p. 562–581. 2009, Carbon and sulfur isotopic anomalies across 10.1111/j.1365-3091.1992.tb02122.x. Veizer, J., Ala, D., Azmy, K., Bruckschen, P., Buhl, D., Bruhn, the Ordovician-Silurian boundary on the Yangtze Plat- Sepkoski, J.J., 1981, A factor analytic description of the F., Carden, G.A.F., Diener, A., Ebneth, S., Godderis, Y., form, South China: Palaeogeography, Palaeoclimatol- Phanerozoic marine fossil record: Paleobiology, v. 7, Jasper, T., Korte, C., Pawellek, F., Podlaha, O.G., and ogy, Palaeoecology, v. 274, p. 32–39, doi: 10.1016/ p. 36–53. Strauss, H., 1999, 87Sr/86Sr, I13C and I18O evolution j.palaeo.2008.12.016. Shackleton, N.J., 1967, Oxygen isotope analyses and Pleisto- of Phanerozoic seawater: Chemical Geology, v. 161, Young, S.A., Saltzman, M.R., Ausich, W.I., Desrochers, cene temperatures re-assessed: Nature, v. 215, p. 15–17, p. 59–88, doi: 10.1016/S0009-2541(99)00081-9. A., and Kaljo, D., 2010, Did changes in atmospheric

doi: 10.1038/215015a0. Waldron, J.W.F., Anderson, S.D., Cawood, P.A., Goodwin, CO2 coincide with latest Ordovician glacial-intergla- Sheehan, P.M., 2001, The Late Ordovician mass extinction: L.B., Hall, J., Jamieson, R.A., Palmer, S.E., Stockmal, cial cycles?: Palaeogeography, Palaeoclimatology, Annual Review of Earth and Planetary Sciences, v. 29, G.S., and Williams, P.F., 1998, Evolution of the Appala- Palaeoecology, v. 296, p. 376–388, doi: 10.1016/ p. 331–364, doi: 10.1146/annurev.earth.29.1.331. chian Laurentian margin: Lithoprobe results in western j.palaeo.2010.02.033. Sigman, D.M., and Boyle, E.A., 2000, Glacial/interglacial Newfoundland: Canadian Journal of Earth Sciences, Zhang, T., Shen, Y., Zhan, R., Shen, S., and Chen, X., 2009, variations in atmospheric carbon dioxide: Nature, v. 35, p. 1271–1287, doi: 10.1139/cjes-35-11-1271. Large perturbations of the carbon and sulfur cycle v. 407, no. 6806, p. 859–869, doi: 10.1038/35038000. Wang, K., Orth, C.J., Attrep, M., Jr., Chatterton, B.D.E., associated with the Late Ordovician mass extinction in SouÀ ane, A., and Achab, A., 2000, Chitinozoan zonation of Wang, X., and Li, J., 1993, The great latest Ordovi- South China: Geology, v. 37, p. 299–302, doi: 10.1130/ the Late Ordovician and the Early Silurian of the Island cian extinction on the South China plate: Chemostrati- G25477A.1. of Anticosti, Québec, Canada: Review of Palaeobotany graphic studies of the Ordovician-Silurian boundary and Palynology, v. 109, p. 85–111, doi: 10.1016/ interval on the Yangtze Platform: Palaeogeography, SCIENCE EDITOR: CHRISTIAN KOEBERL S0034-6667(99)00044-5. Palaeoclimatology, Palaeoecology, v. 104, p. 61–79, ASSOCIATE EDITOR: MAYA ELRICK Swart, P.K., and Eberli, G.P., 2005, The nature of the I13C of doi: 10.1016/0031-0182(93)90120-8. periplatform sediments: Implications for stratigraphy Wang, K., Chatterton, B.D.E., and Wang, Y., 1997, An organic MANUSCRIPT RECEIVED 26 APRIL 2010 and the global carbon cycle: Sedimentary Geology, carbon isotope record of Late Ordovician to Early Silurian REVISED MANUSCRIPT RECEIVED 1 NOVEMBER 2010 v. 175, p. 115–129, doi: 10.1016/j.sedgeo.2004.12.029. marine sedimentary rocks, Yangtze Sea, South China: MANUSCRIPT ACCEPTED 9 NOVEMBER 2010

Underwood, C.J., Crowley, S., Marshall, J.D., and Brench- Implications for CO2 changes during the Hirnantian gla- ley, P.J., 1997, High-resolution carbon isotope stra- ciation: Palaeogeography, Palaeoclimatology, Palaeo- Printed in the USA

1664 Geological Society of America Bulletin, July/August 2011