Late Ordovician conodont biostratigraphy and redefinition of the age of ...

619

Late Ordovician conodont biostratigraphy and redefinition of the age of oil shale intervals on Southampton Island1 Shunxin Zhang

Abstract: Southampton Island was located on the north margin of the Hudson Bay Basin, one of the largest Paleozoic sedimentary basins in North America. The Upper Ordovician sequence on the island includes the Bad Cache Rapids and Churchill River groups and the Red Head Rapids Formation; the latter contains three oil shale intervals in its lower part, which may have significant potential as hydrocarbon source rocks, but their precise biostratigraphic position remains unclear. Numerous conodont elements from 14 localities allow establishment of four interval zones throughout the Upper Ordovician on Southampton Island: the Belodina confluens and Pseudobelodina v. vulgaris zones in the Bad Cache Rapids Group, which are correlative to the upper Edenian – lowest Richmondian Stage; the Amorphognathus ordovicicus Zone from the uppermost Bad Cache Rapids Group to the top of Churchill River Group to the lower Richmondian Stage; and the Rhipidognathus symmetricus Zone in the Red Head Rapids Formation to the upper Richmondian Stage. The oil shale intervals in the lower Red Head Rapids Formation exposed at Cape Donavan, Sixteen Mile Brook, and Boas River on Southampton Island are correlated with the lower R. symmetricus Zone of the upper Richmondian Stage, not the Maysvillian–Richmondian Stage as previous interpreted. The ‘‘Boas River shale’’ and ‘‘Sixteen Mile Brook shale’’ biostratigraphically and lithostratigraphically equate to the Cape Donovan lower and middle oil shale intervals, respectively. The conodont data and continuous sections across the boundaries between the different lithostratigraphic units have proved the presence of Maysvillian strata on Southampton Island, rather than absence as previously interpreted. Re´sume´ : L’ıˆle de Southampton e´tait situeé sur la bordure nord du bassin de la baie d’Hudson, l’un des plus grands bassins se´dimentaires paleózoı¨ques d’Ame´rique du Nord. La se´quence d’Ordovicien supe´rieur sur l’ıˆle comprend les groupes de Bad Cache Rapids et de Churchill River ainsi que la Formation Red Head Rapids; cette dernie`re contient trois intervalles de shales bitumineux dans sa partie infe´rieure, ce qui pourrait constituer un potentiel important en tant que roches sources d’hydrocarbures; cependant, leur position biostratigraphique prećise demeure incertaine. De nombreux e´le´ments de conodontes provenant de quatorze localite´s permettent d’e´tablir quatre zones d’intervalles a` travers l’Ordovicien supe´rieur sur l’ıˆle de Southampton. Les zones a` Belodina confluens et a` Pseudobelodina v. vulgaris dans le Groupe de Bad Cache Rapids sont corre´leés a` l’e´tage E´denien supe´rieur – Richmondien basal; la zone a` Amorphognathus ordovicicus du sommet du Groupe Bad Cache Rapids situe´ au sommet du Groupe Churchill River a` l’e´tage Richmondien infe´rieur et la zone a` Rhipidognathus symmetricus dans la Formation de Red Head Rapids a` l’e´tage Richmondien supe´rieur. Les intervalles de shale bitumineux dans la Formation de Red Head Rapids infe´rieure affleurant a` Cape Donavan, au ruisseau Sixteen Mile et a` la rivie`re Boas sur l’ıˆle de Southampton sont corre´le´s avec la zone a` R. symmetricus infe´rieure de l’e´tage Richmondien supe´rieur et non a` l’e´tage Maysvillien–Richmondien tel qu’interpre´te´ ante´rieurement. Le « shale de Boas River » et le « shale de Sixteen Mile Brooke » constituent respectivement les e´quivalents biostratigraphique et lithostratigraphique des intervalles de shales bitumineux de Cape Donovan infe´rieur et moyen. Les donneés sur les conodontes et les coupes geólogiques a` travers les limites entre les diffe´rentes unite´s lithostratigraphiques ont confirme´ la pre´sence de strates maysvilliennes sur l’ıˆle de Southampton, plutoˆt que leur absence tel qu’interpre´te´ ante´rieurement. [Traduit par la Re´daction]

Introduction The Hudson Bay Basin, one of the largest Paleozoic sediReceived 14 June 2010. Accepted 25 October 2010. Published on the NRC Research Press Web site at cjes.nrc.ca on 31 January 2011. Paper handled by Associate Editor J. Jin. S. Zhang. Canada – Nunavut Geoscience Office, P.O. Box 2319, 626 Tumit Plaza, Suite 202, Iqaluit, NU X0A 0H0, Canada (e-mail: [email protected]). 1Earth

Science Sector Contribution 20100054.

Can. J. Earth Sci. 48: 619–643 (2011)

mentary basins in North America, was developed approximately in the central part of the North American Craton with Southampton Island on its north margin. The lower part of the basin’s Paleozoic succession comprises the Upper Ordovician Bad Cache Rapids and Churchill River groups and Red Head Rapids Formation (Fig. 1), consisting mainly of carbonate with minor shale. Zhang and Barnes (2007) first described and illustrated the conodonts from well cuttings and cores from the Hudson Bay Lowlands and offshore areas (southern and central parts of the Hudson Bay Basin), and then they correlated this sequence with the Edenian– Richmondian, but the Maysvillian-age sediments were interpreted to be questionably missing (Fig. 1).

doi:10.1139/E10-089

Published by NRC Research Press

620

Can. J. Earth Sci. Vol. 48, 2011

Fig. 1. Upper Ordovician stratigraphy on Southampton Island and its position in Late Ordovician time framework. Correlation of the base of Edenian to the base of confluens and superbus zones is adopted from Bergstro¨m et al. (2009). GSSP, global stratotype section and point. Stages: Ch, Chatfieldian; Ed, Edenian; Ma, Maysvillian; Ri, Richmondian; Ga, Gamachian. Fm, Formation; Gp, Group.

Within the Upper Ordovician sequence, two oil shale intervals, ‘‘Sixteen Mile Brook shale’’ (Nelson and Johnson 1966) and ‘‘Boas River shale’’ (Sanford in Heywood and Sanford 1976), were discovered on Southampton Island. The former was interpreted as being stratigraphically either on the top of Churchill River Group (Nelson and Johnson 1966, 1976; Nelson 1981) or in the Red Head Rapids Formation (Sanford in Heywood and Sanford 1976), and the latter as being between the Bad Cache Rapid and Churchill River groups (Sanford and Norris 1973; Sanford in Heywood and Sanford 1976; Ludvigsen 1979; Dewing et al. 1987; Sanford 1987; McCracken and Nowlan 1989; Sanford and Grant 1990, 1998, 2000; Norris 1993; Hamblin 2008). As the oil shale intervals found at Sixteen Mile and Boas River have been considered as potential petroleum source rocks (Macauley 1986; Macauley et al. 1990), debate about the number of oil shale intervals and their stratigraphic position and relative age has lasted over forty years. Discovery of three oil shale intervals within the lower Red Head Rapids Formation in the Cape Donovan area on Southampton Island (Zhang 2008) helped to settle a longlasting debate and proved that the ‘‘Boas River shale’’ is stratigraphically equivalent to the Cape Donovan lower oil shale interval and the ‘‘Sixteen Mile Brook shale’’ equivalent to the Cape Donovan middle or upper oil shale interval. However, the relative ages of the three oil shale intervals still remain unclear, although McCracken and Nowlan (1989) reported the presence of Richmondian and Maysvillian conodonts from ‘‘Sixteen Mile Brook shale’’ and ‘‘Boas River shale’’, respectively. All three oil shale intervals on Southampton Island contain extremely high total organic carbon (TOC) with the highest 34%, and the gamma logs suggest their presence in the Hudson Bay offshore area (Zhang 2008). The potential

of the three oil shale intervals as hydrocarbon source rocks is evident, which necessitates a detailed study to determine their correct position in the Upper Ordovician biostratigraphic framework. This paper will report on the conodonts from outcrops covering the entire Upper Ordovician sequence on Southampton Island and provide an interpretation of relative ages for each lithostratigraphic unit, and particularly, an ascription of relative age to the three oil shale intervals.

Upper Ordovician stratigraphy, sections, and conodont samples Paleozoic rocks, including the Upper Ordovician and Lower Silurian units, are distributed on the southern and western parts of Southampton Island. The Upper Ordovician rocks are discontinuously exposed along the rivers and creeks in the central and southeastern parts of Southampton Island (Fig. 2). Detailed information for each Upper Ordovician stratigraphic unit was provided by Zhang (2008). All the localities and sections (Figs. 2–16) discussed in subsequent sections herein were visited and measured by the author during the field season in 2007; their geographical coordinates are summarized in Table 1. Bad Cache Rapids Group The Bad Cache Rapids Group unconformably overlies the Precambrian basement. It includes two lithological units, a thin basal clastic rock and a dark grey or brownish grey fossiliferous limestone. The limestone, which is the dominant unit of the group, is rich in fossil corals, gastropods, nautiloids, algae, and crinoids. The basal clastic rock and its unconformable contact with the Precambrian basement are well exposed at localities 10 and 11 in the Duke of York Published by NRC Research Press

Zhang

621

Fig. 2. Simplified geological map of Southampton Island. Modified from Zhang (2008) with localities where stratigraphic sections were measured and samples were collected. The white and black (with oil shale intervals) dots represent outcrops where sections were measured: 1, Bad Cache Rapids Group along Fossil Creek; 2–2’, Bad Cache Rapids Group along Rocky Brook; 3, transition between Churchill River Group and Red Head Rapids Formation at unnamed gully; 4, ‘‘bioherms’’; 5, middle oil shale interval, unit 1, Red Head Rapids Formation at Sixteen Mile Brook; 6, lower oil shale interval, unit 1, Red Head Rapids Formation at Boas River; 7, Churchill River Group at Boas River; 8–9, units 2 and 3, Red Head Rapids Formation near junction between Cleveland River and Tungalik Creek; 10–11, Bad Cache Rapids and Churchill River groups along two creeks southeast of Duke of York Bay; 12, Bad Cache Rapids and Churchill River groups in Cape Donovan; 13, units 1–3, Red Head Rapids Formation in Cape Donovan; 14, unit 4, Red Head Rapids Formation and Ordovician– Silurian boundary section in Cape Donovan. See Table 1 for the geographical coordinates of each locality.

Bay area (Figs. 2, 3, 5) and at locality 2 on the upper reaches of Rocky Brook (Figs. 2, 3, 4.1, 6). A nearly complete section of the group was measured at locality 10 in the Duke of York Bay area having a thickness of about 65 m

(Figs. 3, 5). This section at locality 10 was previously recognized as the Churchill River Group by Sanford (in Heywood and Sanford 1976). Within the Bad Cache Rapids Group, a total of 105 conoPublished by NRC Research Press

622


Fig. 3. Upper Ordovician lithostratigraphic units and their correlation among measured sections on Southampton Island. See Fig. 2 for section localities and Table 1 for their geographical coordinates.

dont samples were collected at localities 10 and 11 in the Duke of York Bay area (Figs. 3, 5, 8), locality 2–2’ along Rocky Brook (Figs. 3, 6), locality 1 along Fossil Creek (Figs. 3, 7), and locality 12 in the Cape Donovan area (Figs. 3, 9). The detailed sampling is provided in Table 1. Churchill River Group The Churchill River Group conformably lies on the Bad Cache Rapids Group. It is composed of greenish grey or greyish brown argillaceous limestone. To compare with the underlying Bad Cache Rapids Group, it obviously lacks macrofossils. Sanford (in Heywood and Sanford 1976) interpreted that the ‘‘Boas River shale’’ is a thin unit between the Bad Cache Rapids and Churchill River groups, but this shale interval was not found in most areas in Hudson Bay Basin; therefore, a disconformity between the Bad Cache Rapids and Churchill River groups has been interpreted or illustrated by previous authors, e.g., Norris (1993), Sanford in Heywood and Sanford (1976), and others. This disconformity has been demonstrated as representing a Late Ordovician (Maysvillian) hiatus, or subaerial erosion by previous authors (e.g., Sanford and Norris 1973; Sanford in Heywood and Sanford 1976; Ludvigsen 1979; Dewing et al. 1987; McCracken and Nowlan 1989; Norris 1993). However, a new finding was declared (Sanford and Grant 1990), a 9.8 m of black to dark brown bituminous limestone representative of the ‘‘Boas River shale’’ from a borehole in the Hudson Bay Lowlands. This resulted in a new interpretation that this oil shale interval is conformable and gradational with the Bad Cache Rapids Group below and Churchill River Group above (Sanford and Grant 1990, 1998, 2000; Hamblin 2008). Thus, two questions are raised here: (1) What is the true relationship between Bad Cache Rapids and Churchill River groups? and (2) is there any third lithostratigraphic unit between the two groups? Three sections at localities 10, 11, and 12 (Figs. 3, 5, 8, 9) in the Duke of York Bay and Cape Donovan areas contain the boundary interval between the two groups. It is well exposed at locality 12 in the Cape Donovan area (Figs. 4.2, 9), where the rocks below and above the boundary can be easily differentiated: the thick layered dark grey limestone making up the top of the Bad Cache Rapids Group, and the thin layered dark greenish grey argillaceous dolomitic limestone forming the base of the Churchill River Group. The latter is more easily eroded than the former; therefore, it always forms ‘‘stair step’’ topography (Fig. 4.2). However, no evidence was found to prove the presence of a hiatus or subaerial erosion, and there is no oil shale interval existing between these two groups. No section containing the entire Churchill River Group was found. Due to the limited outcrops, the thickness of the group is unknown on Southampton Island. Within the Churchill River Group, a total of 24 samples were collected at localities 10 and 11 in the Duke of York Bay area (Figs. 5, 11), locality 12 in the Cape Donovan area (Fig. 9), Published by NRC Research Press

Zhang

623

Fig. 4. Outcrops with boundaries between Precambrian and Phanerozoic and between different Upper Ordovician stratigraphic units. (4.1) boundary between Precambrian basement and Bad Cache Rapids Group at locality 2; (4.2) boundary between Bad Cache Rapids and Churchill River groups at locality 12, with an arrow pointing to the boundary; (4.3) transition from Churchill River Group to Red Head Rapids Formation at locality 3; (4.4) close view of shaly dolomitic limestone of lowest Red Head Rapids Formation. Tape measure for scale. Some sample locations are shown (z numbers). See Table 1 for their geographical coordinates.

locality 7 on the upper reaches of Boas River (Fig. 10), and locality 3 in an unnamed gully near the coast between Rocky Brook and Sixteen Mile Brook (Fig. 11). The detailed sampling is provided in Table 1. Red Head Rapids Formation The Red Head Rapids Formation conformably overlies the Churchill River Group. It was divided into three units by Sanford (in Heywood and Sanford 1976). In addition to Sanford’s three units, Zhang (2008) reported a breccia unit between Sanford’s units 1 and 2; thus, the formation is herein divided into four units, including unit 1, thin-layered and laminated argillaceous dolomitic limestone interbedded with three oil shale intervals (29 m); unit 2, massive breccia dolostone–limestone (10–15 m); unit 3, thick, massive biostromal dolostone and limestone (>10 m); and unit 4, thinbedded dolomitic limestone with bioherms (6–7 m). The three oil shale intervals in unit 1 were informally named lower, middle, and upper oil shale intervals (Zhang 2008).

Nelson and Johnson’s (1966) interpreted that ‘‘Sixteen Mile Brook shale’’ is stratigraphically located on the top of Churchill River Group, which was ascribed as being positioned within the Red Head Rapids Formation (Sanford in Heywood and Sanford 1976), and then as being correlated with middle or upper oil shale interval of the unit 1, Red Head Rapids Formation (Zhang 2008). In fact, the boundary between the Churchill River Group and Red Head Rapids Formation is transitional; practically, the base of the Red Head Rapids Formation is marked by a 15 cm dark brown, shaly dolomitic limestone at locality 3 (Figs. 4.3, 4.4), an unnamed gully near the coast between Rocky Brook and Sixteen Mile Brook, by the author during the field study in 2007. Within the Red Head Rapids Formation, a total of 140 samples were collected at locality 3, an unnamed small gully near coast between Sixteen Mile Brook and Rocky Brook (Fig. 11); locality 13, i.e., Zhang’s (2008) three oil shale type localities and locality 14 in the Cape Donovan area Published by NRC Research Press

624

Fig. 5. Conodont distribution in Bad Cache Rapids and lower Churchill River groups at locality 10 in Duke of York Bay area. See Fig. 3 for lithology and Table S82 for numerical data and Geological Survey of Canada (GSC) curatorial numbers.



Fig. 6. Conodont distribution in Bad Cache Rapids Group at locality 2–2’ along the Rocky Brook. The strata below and above the breaks in stratigraphic column represent roughly lower and upper stratigraphic relationship, but different geographical coordinates from north to south along Rocky Brook; the longer scale bar is for the lower part of the section from sample Z006-01-11, and the shorter scale bar is for rest of the section. See Fig. 3 for lithology and Table S22 for numerical data and GSC curatorial numbers. * indicates a barren sample.

Zhang 625


626


Fig. 7. Conodont distribution in middle Bad Cache Rapids Group at locality 1 along Fossil Creek. The strata below and above the breaks in stratigraphic column represent roughly lower and upper stratigraphic relationship, but different geographical coordinates from north to south along Fossil Creek. See Fig. 3 for lithology and Table S12 for numerical data and GSC curatorial numbers.

(Fig. 12); locality 6, i.e., Sanford’s (in Heywood and Sanford 1976) ‘‘Boas River shale’’ type locality (Fig. 13); locality 5, i.e., Nelson and Johnson’s (1966) ‘‘Sixteen Mile Brook shale’’ type locality (Fig. 14); localities 8 and 9 along the Tungalik Creek (Fig. 15); and locality 4, an isolated bioherm unit near the coast between Rocky Brook and Sixteen Mile Brook (Fig. 16). The detailed sampling is provided in Table 1.

Late Ordovician conodont biostratigraphy Zhang and Barnes (2007) defined three conodont zones within the Upper Ordovician sequence using conodonts from continuous cores and well cuttings from six exploration wells in the Hudson Bay Lowlands and offshore areas. The zones include Belodina confluens, Amorphognathus ordovicicus, and Rhipidognathus symmetricus interval zones. Published by NRC Research Press

Zhang

627

Fig. 8. Conodont distribution in upper Bad Cache Rapids and lower Churchill River groups at locality 11 in Duke of York Bay area. See Fig. 3 for lithology and Table S92 for numerical data and GSC curatorial numbers.

These zones were tentatively correlated with the Bad Cache Rapids Group, middle–upper Churchill River Group, and Red Head Rapids Formation, respectively. An Edenian age was interpreted for the Bad Cache Rapids Group and Richmondian age for the Churchill River Group and Red Head Rapids Formation (Zhang and Barnes 2007). However, because of the limited material, wide sampling intervals varying from 10 to 120 feet (1 foot = 0.3048 m), and five out of six exploration wells lacking any cores for observation, Zhang and Barnes (2007) were indeterminate of (1) the precise age of each zone, (2) whether the strata representing the 2 Supplementary

Maysvillian Stage exist in Hudson Bay Basin, and (3) most importantly, the precise age of the oil shale intervals. This present report is based on a study of conodonts from 269 conodont-bearing samples from 14 localities on Southampton Island. In total, about 15 000 conodont specimens plus numerous uncounted specimens of Panderodus Ethington and Drepanoistodus Lindstro¨m were collected, among which 36 species representing 23 genera were identified, with 1 element indet. (Figs. 17–20). The numerical conodont distribution data are presented in Tables S1–S112, and the conodont stratigraphical distributions at the 14 localities are

data tables are available with the article through the CJES Web site. Published by NRC Research Press

628


Fig. 9. Conodont distribution in upper Bad Cache Rapids and lower Churchill River groups at locality 12 in Cape Donovan area. See Fig. 3 for lithology and Table S102 for numerical data and GSC curatorial numbers.

Fig. 10. Conodont distribution in lower-middle Churchill River Group at locality 7 on the upper reaches of Boas River. See Fig. 3 for lithology and Table S72 for numerical data and GSC curatorial numbers.


Zhang

629

Fig. 11. Conodont distribution in upper Churchill River Group and lower Red Head Rapids Formation at locality 3, an unnamed gully near coast between Rocky Brook and Sixteen Mile Brook. See Fig. 3 for lithology and Table S32 for numerical data and GSC curatorial numbers. * indicates a barren sample.

shown in Figs. 5–16. Based on these data, four interval zones (abbreviated as ‘‘zones below’’) are recognized, and the 14 sections are correlated biostratigraphically (Fig. 3). Definition of conodont zones Belodina confluens Zone (revised) The base of the Belodina confluens Zone is defined by the lowest occurrence of the nominate species (Figs. 18.14– 18.17), which is in sample Z006-01A, 10 cm above the basal sandstone of the Bad Cache Rapids Group at locality 2, on the upper reaches of Rocky Brook (Figs. 4.1, 6). The top of the zone is marked by the base of the overlying Pseudobelodina v. vulgaris Zone. Pseudobelodina v. vulgaris Zone (new) The base of the P. v. vulgaris Zone is defined by the lowest occurrence of the nominate species (Figs. 18.25–18.28), which is in sample Z011-06, middle Bad Cache Rapids Group at locality 2, on the middle reaches of Rocky Brook (Fig. 6). The top of the zone is marked by the base of the overlying Amorphognathus ordovicicus Zone. Amorphognathus ordovicicus Zone (revised) The base of the A. ordovicicus Zone is defined by the

lowest occurrence of the nominate species (Figs. 19.17– 19.23), which is in sample Z030-04, uppermost Bad Cache Rapids Group, about 20–30 cm below the boundary between the Bad Cache Rapids and Churchill River groups at locality 12 in the Cape Donovan area (Figs. 4.2, 9). The top of the zone is marked by the base of the overlying Rhipidognathus symmetricus Zone. Rhipidognathus symmetricus Zone (revised) The base of the R. symmetricus Zone is defined by the lowest occurrence of the nominate species (Figs. 20.1– 20.9), which is in sample Z034-03, at the base of unit 1, Red Head Rapids Formation at locality 3, an unnamed gully between Rocky Brook and Sixteen Mile Brook near the coast (Figs. 4.3, 4.4, 11). The top of the zone is marked by the unconformity between the Upper Ordovician Red Head Rapids and Lower Silurian Severn River formations. Stratigraphic distribution and age of conodont zones Belodina confluens Zone The Belodina confluens Zone has been recognized in the upper Chatfieldian and lower Edenian in the North American Midcontinent Province (e.g., Sweet 1984; Harris et al. 1995; Webby et al. 2004); however, the base of the Edenian Published by NRC Research Press

Fig. 12. Conodont distribution in Red Head Rapids Formation at localities 13 and 14 in Cape Donovan area. The strata below and above the break in stratigraphic column represent roughly lower and upper stratigraphic relationship but different geographic coordinates. See Fig. 3 for lithology and Table S112 for numerical data and GSC curatorial numbers. * indicates a barren sample.

630 Can. J. Earth Sci. Vol. 48, 2011


Zhang

631

Fig. 13. Conodont distribution in an interval including lower oil shale interval and strata below and above within unit 1, Red Head Rapids Formation at locality 6 on the upper reaches of Boas River. See Fig. 3 for lithology and Table S62 for numerical data and GSC curatorial numbers. * indicates a barren sample.

has not been closely defined faunally in the type area — Cincinnati region. Recently, Bergstro¨m et al. (2009) stated that it is just above the base of the B. confluens Zone in the North American Midcontinent Province. The boundary between Edenian and Maysvillian in the Cincinnati region falls in the middle of the Oulodus velicuspis Zone (e.g., Sweet 1984; Harris et al. 1995; Webby et al. 2004). Thus, the B. confluens Zone only represents the lower Edenian in the North American Midcontinent Province, based on Bergstro¨m et al. (2009). The Belodina confluens Zone was previously recognized in the Bad Cache Rapids Group in the Hudson Bay Lowland and offshore area (Zhang and Barnes 2007) based on the well materials; because of insufficient conodont data and inadequate analysis of the entire conodont fauna, the upper boundary of the zone was not confirmed and the zone was dated as Edenian.

The conodonts collected from close intervals (most