INTRODUCTION AND REGIONAL SETTING

1 The Magdalen Islands lie near the centre of the Gulf of St. Lawrence in the western part of the Maritimes Basin (Fig. 1). The western Maritimes Basin covers an area of more than 200 000 square kilometres and has a sedimentary fill ranging from mid-Devonian to Early Permian in age. Total thickness of the basin fill exceeds 12 km in the deepest central portions of the basin, as determined from seismic reflection data (Durling and Marillier 1993).

Display large image of Figure 1

2 The Windsor Group comprises a regional marker sequence within the Carboniferous succession. Where typically developed, the group is characterized by relatively thin marine carbonate rock units which were synchronously deposited over large areas, intercalated with evaporite units and non-marine detrital clastic rocks (Giles 1981). With a long history of biostratigraphic research and a distinctive pattern of cyclic marine and non-marine sedimentation, the Windsor Group provides an excellent vehicle for basin reconstruction. Bedrock exposures on the Magdalen Islands provide a unique onshore opportunity to assess the stratigraphic succession in this part of the Maritimes Basin. Of particular interest in the Magdalen Islands succession are volcanic rocks, principally of basaltic composition, which are interbedded with Mississippian marine strata and which are known to occur in the Windsor Group only at this location.

3 The purpose of the present investigation is 1) to document the lithostratigraphy of the Windsor Group on the Magdalen Islands, 2) to assess biostratigraphic constraints on the position of the basaltic rocks within the Windsor Group, applying available micropaleontological and palynological data in order to better estimate the timing of basaltic extrusion, and 3) to place the succession within the regional stratigraphic framework of the Windsor Group.

GEOLOGICAL SETTING

4 The Magdalen Islands archipelago consists of a group of islands situated in the central portion of the Gulf of St Lawrence (Fig. 1). Carboniferous to Permian(?) bedrock underlies the islands, six of which are interconnected by modern sand spits and tombolos (Fig. 2). Windsor Group strata crop out on six of the islands, typically in fault-bound blocks, flanked by gently-dipping Permian(?) sandstone. The general structural configuration for the Windsor Group outcrop areas is best seen on the three largest islands, Havre aux Maisons, Havre Aubert, and Cap aux Meules (Fig. 3). There, major faults trending approximately east-west bound anticlinal horsts composed of volcanic and associated sedimentary rocks. These fault-bound structures are flanked by gently dipping strata of presumed Permian age, which unconformably overlie the Windsor Group at the single observed locality where the boundary is not a fault.

5 Windsor Group coastal exposures within the archipelago are typically spectacular and reasonably accessible, but are also severely deformed. Salt diapirism has been suggested as the mechanism responsible for bringing Windsor Group strata to the present surface and may be largely responsible for the complex, intensely sheared and folded character of the Mississippian succession (Brisebois 1979, 1981).

Display large image of Figure 2

Display large image of Figure 3

PREVIOUS WORK AND THE DEVELOPMENT OF STRATIGRAPHIC NOMENCLATURE

6 The strategic location of the Magdalen Islands from the perspective of regional Carboniferous studies has been appreciated since the early work of Clarke (1911) and Alcock (1941). Bell (1948, unpublished informal Geological Survey of Canada report, 15 pages) was the first to attempt to establish detailed stratigraphic relationships, relying heavily on his experience with the Windsor Group elsewhere in Atlantic Canada. He recognized, mainly using macrofaunal biostratigraphic evidence, that marine fossils of both the lower and upper biostratigraphic zones of the Windsor Group (Bell 1929) were represented in limestone beds exposed on the islands, assigning the terms `Adele’ and `Coffin Island’ beds, respectively, to the corresponding local lithostratigraphic subdivisions. Bell believed that the volcanic rocks are associated entirely with the lower `Adele beds’. Bell assigned thick sandstones unconformably overlying the Windsor Group to the Alright Formation, a term no longer in use, and correlated the latter formation with rocks of the Canso Group elsewhere in Atlantic Canada. The term Canso also is no longer in use and has been replaced by the Mabou Group (see Ryan et al. 1991).

7 Sanschagrin (1964) produced the first modern geological map of the Magdalen Islands, building on the previous stratigraphic work of Bell and suggesting new terminology for the entire succession (Fig. 4). Sanschagrin assigned Windsor Group volcanic and associated sedimentary rocks to a single formation termed the Havre aux Maisons Formation. He named a lower Cap Adèle Member to which he assigned the beds of Bell’s `Adele Beds’, and an upper Bassin aux Huîtres Member (Fig. 4) which corresponded to Bell’s `Coffin Island Beds’. Like Bell, Sanschagrin interpreted the volcanic rocks to lie entirely within the Adèle Member in the lower part of his Havre aux Maisons Formation. Sanschagrin (1964) reported volcanic rocks of his Havre aux Maisons Formation at a single additional locality on Île du Corps Morts, a small island approximately 16 km west of Île du Havre Aubert (Fig. 3, inset map 1). These exposures were not included in the present study as the island can be safely accessed only during exceptionally calm conditions. Above the Havre aux Maisons Formation, Sanschagrin named the Cap aux Meules Formation, replacing Bell’s informal Alright `formation’. He was the first to postulate that the sandstone above the Havre aux Maisons Formation might be as young as Permian in age, while still suggesting regional assignment to the Canso (presently Mabou) Group.

Display large image of Figure 4

8 Brisebois (1979, 1981) prepared geological maps of the archipelago, significantly modifying the stratigraphic framework of Sanschagrin (1964). He retained the name Havre aux Maisons Formation but restricted the formation to the lower portion of Sanschagrin’s earlier defined formation (Fig. 4). Brisebois (1979, 1981) described both volcaniclastic rocks and basaltic flows in his redefined Havre aux Maisons Formation, associated with gypsum, mudstone, and thin marine carbonate rocks. Members of the Havre aux Maison Formation established by Sanschagrin (1964) were abandoned. For the upper part of Sanschagrin’s Havre aux Maisons Formation, Brisebois (1979, 1981) erected the Cap au Diable Formation, characterized by volcanic rocks represented principally by basaltic flows. He considered the Cap au Diable Formation to be laterally equivalent to Sanschagrin’s Bassin aux Huîtres Member (Bell’s Coffin Island Beds, Fig. 4), and to represent the upper part of the Windsor Group on the southern islands of the archipelago. The contact between Brisebois’ Havre aux Maisons and Cap au Diable formations was reported to be transitional, and placed immediately above the highest interval of mudstone in the Havre aux Maisons Formation. Although Bell (1948) speculated on possible domal salt structures beneath the Magdalen Islands, Brisebois (1979, 1981) first documented thick salt deposits revealed by deep drilling. These beds were not formally named by Brisebois, who assigned them to the lower Windsor Group, lying beneath his Havre aux Maisons Formation. Several investigations of Windsor Group rocks exposed on the Magdalen Islands have relied heavily on the stratigraphy erected by Brisebois (1981) (see, for example, Barr et al. 1985; Plint and von Bitter 1986; Howie 1988).

9 Above the Windsor Group, Brisebois (1981) subdivided the Cap aux Meules Formation of Sanschagrin (1964) into a lower Étang du Nord Member, and an upper Étang des Caps Member (Fig. 4). The Étang du Nord Member comprises sandstone, siltstone, and mudstone with small amounts of dolomitic limestone. Brisebois (1981, p.30) reported an approximate thickness of 300 m for the member on Île du Havre Aubert. The overlying Étang des Caps Member comprises fine-grained and very fine-grained sandstone, characterized by thick beds and large-scale cross-stratification. The thickness of the Étang des Caps Member was estimated at 370 m by Brisebois (1981, p. 35). Brisebois interpreted the Cap aux Meules Formation to be no older than Stephanian, and suggested a possible Permian age, based on tentative correlation with red strata of that age on Prince Edward Island in the southern part of the Gulf of St. Lawrence.

10 The chemical affinities of Magdalen Island basalts and their significance to models for Maritimes Basin evolution have been the focus of several research efforts. Barr et al. (1985) addressed the chemistry and petrography of basaltic rocks of the Windsor Group, reporting possible alkaline affinities, but noted the severely altered nature of the volcanic rocks. More recent study of the Magdalen Island basalts (La Flèche et al. 1998) addressed their geochemistry and implications for their origin, and followed Brisebois in assigning the basalts to the highest part of the Windsor Group. La Flèche et al. (1998) reported the presence of geochemically heterogeneous alkalic and tholeiitic basalts. Pe-Piper and Piper (1998) included several samples of Magdalen Island basalt in an assessment of isotopic evidence for mantle and lower crustal sources for Devonian-Carboniferous rocks of the western Maritimes Basin. They concluded that only relatively young, Visean-Westphalian alkalic magmas in this region showed isotopic evidence which might allow input from an asthenospheric plume source. Because of their alteration, these basaltic rocks have not been dated radiometrically. Here, their biostratigraphically constrained stratigraphic position provides some estimate for the timing of basaltic extrusion.

11 Biostratigraphic studies have considerably expanded the first macrofaunal assessment of Bell (1948). Sanschagrin (1964) reported macrofaunas recovered from marine limestone on the Magdalen Islands and followed Bell (1948) in assigning the host strata to lower and upper Windsor Group macrofaunal sub-zones B, C, D, and E (Fig. 5). Brisebois (1979) and Brisebois and Mamet (1974) reported calcareous foraminiferans and algae from carbonate rocks of the Magdalen Islands with emphasis on the upper parts of the succession. More recent biostratigraphic investigations include an assessment of conodont faunas from marine carbonate rocks of the Magdalen Islands (Plint and von Bitter 1986; von Bitter and Plint 1987), and studies of palynomorphs in Windsor Group strata by Utting (1987). Utting (1987), unlike most recent workers in the archipelago, followed the older lithostratigraphic framework established by Sanschagrin (1964).

Display large image of Figure 5

12 Attempts to correlate specific marine carbonate members of the upper Windsor Group on the Magdalen Islands with members recognized in the Windsor Group type area include the combined litho- and biostratigraphic assessment of McCulloch (1973), who studied what he believed to represent a possible correlative of the Kennetcook Limestone Member of Moore (1967; Moore and Ryan 1976, fig. 15), exposed on Île Boudreau. Blakeney (1974) reported on the chemistry and petrography of the most northeasterly exposed carbonate member on Île Boudreau, correlating that carbonate member with the Herbert River Limestone Member (Moore 1967) of the type upper Windsor Group. Blakeney did not address the macrofauna of this carbonate member in detail but alluded to key taxa which suggested correlation with the Herbert River limestone.

13 All stratigraphic assessment of Magdalen Island exposures are hampered by structural complexities, the apparent lack of unequivocal marker horizons, and uncertainties in even local facing directions. In spite of these difficulties, the present work shows that remnants of the original stratigraphic succession can be documented using both litho- and biostratigraphic data.

Display large image of Figure 6

WINDSOR AND MABOU GROUP(?) STRATIGRAPHY

14 Lithostratigraphically, the Windsor Group can be considered regionally in three parts. At the base of the group, a lower carbonate-sulphate-halite sequence represents a single thick cycle of sedimentation. This sequence is marked at the base by the Macumber Formation (limestone) in Nova Scotia and New Brunswick, and the Ship Cove Formation in southwestern Newfoundland. In basin-margin settings, equivalent strata are typically biohermal and are assigned to the Gays River Formation or local equivalents. Thick anhydrite and salt deposits overlying the Macumber Formation complete this initial transgressive-regressive depositional episode. These rock units are here assigned collectively to the lower Windsor Group (lithostratigraphic). The middle portion of the group comprises numerous repeated cycles of marine carbonates with capping evaporites and non-marine sediments. The upper part of the Windsor Group as described by Moore (1967) is dominated by non-marine strata but contains at least eight marine carbonate members with or without capping evaporites.

15 The upper Windsor Group of Moore (1967) coincides with the Upper Windsor Zone of Bell (1929) in regional biostratigraphic terminology. The middle Windsor Group as used in this study corresponds to macrofaunal Subzone B of Bell (1929) and the lower Windsor Group (lithostratigraphic) corresponds to Bell’s macrofaunal Subzone A. Subzones A and B together comprise the Lower Windsor Zone in generally applied biostratigraphic practice. The inter-relationships between Windsor Group biostratigraphic and lithostratigraphic frameworks are illustrated in Figure 5.

16 In this paper, italics are used in the text to identify the lower and upper Windsor Group macrofaunal zones of Bell (1929) where this is necessary for the sake of clarity, and to clearly distinguish these zones from the informal lithostratigraphic designations used herein of lower, middle, and upper Windsor Group. Formal Windsor Group lithostratigraphic units applied on the Magdalen Islands and accepted by the present author include the Havre aux Maisons Formation (sensu Sanschagrin 1964), and the Cap Adèle and Bassin aux Huîtres members of that formation, again as defined by Sanschagrin (1964). Strata equivalent to the lower part of the Mabou Group in Nova Scotia, here tentatively identified through biostratigraphic assessment using palynological data from Utting (1987), are assigned to that position without additional formal lithostratigraphic terminology. They are represented in a single outcrop.

DISCUSSION

34 The evidence for assignment of the volcanic rocks of the Magdalen Islands succession to the Windsor Group is unequivocal. Blocks of basalt incorporated within bedded marine limestone, and more definitively the interstratification of basaltic tuff with fine-grained siltstone and mudstone of the biostratigraphically dated middle Windsor Group succession indicate clearly that volcanism was a part of the depositional history of the Windsor Group in the archipelago. The restriction of volcanic rocks to the middle part of the Windsor Group, i.e., the lower member of the Havre aux Maisons Formation, contrasts with the previous interpretation of Brisebois (1981) and warrants discussion here.

35 Simply stated, Magdalen Island volcanic rocks occur in stratigraphic continuity only with strata assigned here to the Cap Adèle Member. In the Île Boudreau section, where marine strata are largely equivalent to beds of the upper Windsor Group, no volcanic rocks are seen. On Île d’Entrée, volcanic rocks are extensive but are not in stratigraphic contact with fault-bound upper Windsor strata. Furthermore, on that island, beds which appear to be in stratigraphic contact with basalts were assigned biostratigraphically (Plint and von Bitter 1986) to the lower portion of the Havre aux Maisons Formation, the Cap Adèle Member of this study. Although there is no a priori reason to suggest that the volcanic rocks could not occur throughout the Windsor Group succession, data presently available cannot confirm this broader distribution. Basaltic rocks sparsely represented within the diapiric lower Windsor Group salt in boreholes M16 and M22 have uncertain stratigraphic validity. Thus, the writer’s assessment of lithostratigraphic position (Figs. 4 and 6) is the most cautious but the best supported hypothesis, and is in agreement with the original interpretations of both Bell (1948) and Sanschagrin (1964). The rationale for the Cap au Diable Formation of Brisebois (1979, 1981) as a major volcanic-bearing rock unit, equivalent in age to the upper Windsor Group, cannot be supported with present data and that term is here abandoned (Fig. 4). The stratigraphy of the Cap Adèle Member, which is represented by the measured sections of Figure 6, suggests that basalts can be assigned to two principal horizons. Because these sections are not constrained at the top or base by recognized contacts, the member could contain basaltic rocks at additional horizons.

36 The occurrence of basalt within the late Visean succession in the archipelago is unique in rocks of this age in eastern Canada, although basalts have been reported from Tournaisian and Namurian Maritimes Basin strata in New Brunswick (Fyffe and Barr 1986) and from mid-Visean rocks of Cape Breton Island (Barr et al. 1994). Within the Gulf of St. Lawrence, in the broad vicinity of the Magdalen Islands, several deep hydrocarbon exploration wells have penetrated the Windsor Group succession. The Brion Island well, drilled approximately 20 km north of the main archipelago, failed to intersect any volcanic rocks in a relatively thick section of evaporite, shale, and minor carbonate rocks biostratigraphically identified as Windsor Group equivalents (Giles and Utting 2003). The Shell-Amoco Cap Rouge F-52 well, drilled south of the Magdalen Islands in the Gulf of St. Lawrence, drilled a complete section of the Windsor Group (Giles and Utting 2001) without volcanic rocks. The Bradelle L-49 well, drilled west of the Magdalen Islands, intersected only a thin interval of marine carbonate rocks equivalent to the Windsor Group (Giles and Utting 2003), and likewise failed to provide documentation of any basaltic rocks of comparable age.

37 The age of the basaltic rocks is biostratigraphically indicated but not presently substantiated by any radiometric dates. Von Bitter et al. (2006) assigned strata of the middle part of the Windsor Group, including Bell’s B subzone and Cycle 2 of Giles (1981), to the late Asbian substage of George et al. (1976). The correlations suggested here likewise assign the basaltic rocks of the Magdalen Islands Windsor Group to the late Asbian. Time Scale B of Menning et al. (2000), the most recently published attempt to summarize absolute ages for the British Dinantian substages, estimates that the Asbian extends from 331 to 334.5 Ma, which provides some tentative constraint on the absolute age of the Magdalen Island basalt. The late Asbian should be represented by the younger part of the Asbian (total) range, perhaps from 331 to 333 Ma. Basaltic lavas and tuffs of Asbian age have been noted in southwestern Ireland in the Limerick syncline (Somerville et al. 1992). Clays with K-bentonitic compositions containing igneous-derived accessory minerals were noted from eight discrete thin levels within latest Asbian strata in the vicinity of the Askrigg platform in north-central England by Walkden (1979) who also noted that similar K-bentonite also occurs near the top of the Brigantian substage. The presence of Asbian K-bentonitic clay horizons in North Wales was reported by Somerville (1979).

38 Rees et al. (1996) described Asbian basaltic tuffs in a deep borehole at Apedale in central England and linked this volcanic activity to much more widespread volcanism. They attributed this volcanism to active late Asbian and early Brigantian rifting which affected much of northwestern Europe. The Apedale volcanic centre was described as larger than most of comparable age in central England, with tuffaceous deposits ranging from 100s – ~1000 m in thickness, covering an estimated and very modest 100 km². The Magdalen Islands basaltic flows and tuffs are comparable in presently documented (minimum) lateral extent and estimated thickness to Asbian volcanic rocks described in the Apedale borehole. If these rocks represent an analogous centre of Asbian volcanic activity in eastern Canada, the extent of the local basaltic rocks could be likewise very limited. However, other similar centres of volcanic activity could easily escape recognition in the submarine portions of the Maritimes Basin beneath the Gulf of St. Lawrence.

39 Windsor Group rocks on the Magdalen Islands occur in easterly to east-northeasterly trending fault-bound horst blocks (Brisebois 1981). Ryan and Boehner (1994) showed that Windsor Group salt movements in the Cumberland Basin of Nova Scotia often exploited major fault systems, resulting in linear faulted anticlinal structures, cored by Windsor Group salt in the subsurface. A similar mechanism might explain the elongate salt-cored horst blocks of the Magdalen Islands. Pe-Piper and Piper (1998) noted that alkalic magmas of Visean to Westphalian age in the western Maritimes Basin (or the Magdalen Basin) ascended along crustal-scale faults. If the Magdalen Island horst blocks are the surface manifestation of deeper crustal faults, the salt-basalt association would be consistent with igneous intrusion ascending the same deep-seated faults which subsequently focused diapiric salt movements. The salt-basalt association suggests that potential field magnetic and gravity data might be useful in tracing the major structures of the Magdalen Islands offshore into the adjacent Gulf of St. Lawrence.

40 Within the Bassin aux Huîtres Member of the Havre aux Maisons Formation, carbonate members equivalent to both the Herbert River Limestone Member and, less confidently, the Kennetcook (or E1) Limestone Member can be recognized. These members in the type Windsor Group mark the base and top, respectively, of the upper part of the Windsor Group (Moore 1967; Giles and Boehner 1982). Two additional upper Windsor Group carbonate members are exposed on Ile Boudreau, but their lithostratigraphic correlation with type Windsor Group members is not certain. It is of particular importance that both the lowest and the highest carbonate members of the upper Windsor Group appear to be represented in exposures on the Magdalen Islands. Their presence, overlying a thick succession of middle Windsor Group strata also containing a significant number of marine carbonate intervals, suggests that a large and relatively complete portion of the Windsor Group is represented. This situation contrasts with sections in northwestern Nova Scotia and southeastern New Brunswick where upper Windsor Group carbonate beds are absent (Ryan and Boehner 1994) and middle Windsor Group marine bands dramatically reduced in number, possibly represented only in the lowest portion of the middle Windsor. The Magdalen Islands exposures are most comparable with the complete Windsor Group succession documented in the type area and in the Shubenacadie and Musquodoboit Basins of southern Nova Scotia, and throughout Cape Breton Island, in terms of stratigraphic completeness and longevity of marine influence (Fig. 12).

Display large image of Figure 12

CONCLUSIONS

41 In spite of structural complexity, vestiges of the Late Visean Windsor Group succession can be documented on the Magdalen Islands. Lower Windsor Group strata are represented by thick salt deposits which underlie the islands, known only from the subsurface. The middle Windsor Group is represented by a total estimated thickness (reconstructed) approaching 1200 m, comprising gypsum, red and lessor grey siltstone, marine carbonate rocks, and volcanic rocks. The latter include both basaltic flows and basaltic volcaniclastic rocks, with significant thicknesses of basalt concentrated at two distinct levels within the middle Windsor Group. The basaltic rocks are of late Asbian age. Strata of the upper Windsor Group are more limited in their distribution in coastal exposures, but appear to include local representatives of key marine strata of the type upper Windsor Group, including the stratigraphically highest beds. No reliable estimate of total thickness for the upper Windsor Group on the Magdalen Islands can be made. Volcanic rocks cannot be demonstrated to interbed with strata of the upper Windsor Group on the Magdalen Islands. The Cap au Diable Formation, erected as a volcanic-dominated rock unit of the upper Windsor Group on the Magdalen Islands, is therefore abandoned. In its degree of completeness as indicated by the preservation of marine marker horizons, the Magdalen Islands succession compares best with the Windsor Group in southern and eastern Nova Scotia, including Cape Breton Island.