INTRODUCTION

1 Most of Isle Madame and adjacent Petit-de-Grat Island (subsequently referred to collectively as Isle Madame) on the southern tip of Cape Breton Island, Nova Scotia (Fig. 1), consist of Carboniferous sedimentary rocks and pre-Carboniferous “basement” units (Weeks 1954; White and Barr 1998; Giles et al. In press). Geological relationships on Isle Madame are unlike those elsewhere in the region in that the pre-Carboniferous rocks occur in elongate, fault-bounded belts separated by thick, steeply dipping, conglomerate-dominated Carboniferous successions (Fig. 2). The purpose of this study is to investigate the significance of these conglomerate units and their relation to the older rocks in terms of both local and regional structural settings, and to suggest a crustal extension model for their development.

2 The two-fold division of the rocks of Isle Madame into basement and conglomerate-dominated sedimentary successions was recognized by previous workers in the area, beginning with Fletcher (1881), who interpreted the basement rocks as Precambrian and the sedimentary rocks as Devonian. Weeks (1954, 1964) included the basement rocks in the Precambrian Fourchu Group of southeastern Cape Breton Island, although he acknowledged their unusual and apparently exotic character. He described the rocks as fine breccia and tuff, or grey-wacke composed of volcanic material, and mentioned that they have been granitized adjacent to “rhyolite intrusions”. Barr et al. (1982) reported petrographic and chemical data from the Petit-de-Grat pluton, one of the “rhyolite intrusions” of Weeks (1954), located on the southeastern tip of Petit-de-Grat Island. Cormier (1980) reported a Rb-Sr isochron age of 357 ± 11 Ma for the pluton, and Keppie (1982) mentioned a 40Ar/39Ar amphibole cooling age of about 391 Ma for amphibolite from Green Island, located offshore from Isle Madame (Fig. 2). Barr et al. (1992) noted the varied metamorphic and mylonitic rocks in the basement blocks on Isle Madame and Green Island, and suggested that they were derived from the Meguma terrane. White and Barr (1998, 1999) included the metamorphic rocks in a map unit termed the Chedabucto fault complex, and assigned the adjacent mainly conglomeratic rocks to the mid-Devonian Glenkeen Formation. Subsequent mapping and palynology resulted in reassignment of some of the Glenkeen Formation of White and Barr (1998, 1999) to the Horton Group (Giles et al. In press).

Display large image of Figure 1

3 However, in spite of these advances in the understanding of the distribution of Carboniferous and pre-Carboniferous rock units and refinement of their ages, the reason for the fault-bounded belts of metamorphic rocks and Glenkeen Formation and the atypical abundance of conglomerate in the Carboniferous units remained uncertain. The present study is an attempt to resolve these enigmas.

REGIONAL DEVONIAN AND CARBONIFEROUS GEOLOGY

4 The Maritimes Basin was initiated following the Acadian orogeny (ca. 400–390 Ma; Hicks et al. 1999), apparently as a number of half-graben depocentres related to regional extension (e.g., Belt 1968; Hamblin and Rust 1995; Calder 1998; Waldron et al. 2005), although the mechanism of basin formation is uncertain (e.g., Bradley 1982; McCutcheon and Robinson 1987; Murphy and Keppie 1998; Lynch and Tremblay 1994; Lynch 2001). Lynch and Tremblay (1994) and Lynch (2001) suggested that basin development included a Late Devonian regional detachment, which they termed the Margaree detachment.

Display large image of Figure 2

5 The mid to late Devonian part of the basin fill is exposed only at scattered locations in Cape Breton Island and northern mainland Nova Scotia, and generally consists of interbedded volcanic and sedimentary rocks, as exemplified by the Glenkeen Formation in the Guysborough area (Cormier et al. 1995; White and Barr 1999) and the Fisset Brook Formation in Cape Breton Island (e.g., Dunning et al. 2002). These older units are unconformably to conformably overlain by the more voluminous and widespread younger and dominantly sedimentary units of the basin. The latter units are the mainly nonmarine Horton Group (latest Devonian to Tournaisian), overlain unconformably by the shallow marine and nonmarine Windsor Group (middle Visean) and the nonmarine Mabou, Cumberland, and Pictou groups (late Visean to early Permian).

6 Although volcanic rocks occur mainly in the earliest part of the basin fill, they also occur higher in the succession. Gabbroic plutons and related minor basalt flows in southern Cape Breton Island have been dated at 339 ± 2 Ma (U-Pb, zircon and baddeleyite; Barr et al. 1994), and appear to coincide with a regional unconformity between the Horton and Windsor groups (Reynolds et al. 2004). Locally in western Cape Breton Island and elsewhere, evidence for igneous activity during the time of Horton Group deposition is present in the form of mafic and felsic volcanic rocks and mafic dykes (e.g., Barr and Peterson 1998; Pe-Piper and Piper 1998; Dunning et al. 2002). Petrological characteristics of the volcanic rocks are generally consistent with eruption in a continental extensional setting (e.g., Barr et al. 1995; Barr and Peterson 1998; Pe-Piper and Piper 1998).

7 The Horton Group in the Maritimes Basin is commonly regarded as an intra-continental, fluvial to lacustrine basin-fill succession that oversteps terrane boundaries (e.g., Calder 1998 and references therein). In mainland Nova Scotia, the basin-fill successions overlie rocks of both the Meguma and Avalon terranes, which were brought into juxtaposition during the Devonian along the Cobequid-Chedabucto fault zone (Fig. 1). Seismic profiles and sedimentary evidence indicate that Horton Group strata generally occupy local half grabens with varied polarity along this fault and other structures. The depositional succession in each Horton subbasin typically consists of (i) a thick basal coarse-grained alluvial succession that rests unconformably on older “basement” rock, overlain by (ii) a dark-grey, fine-grained lacustrine succession, which is overlain conformably to unconformably by (iii) an upper coarse- to fine-grained alluvial succession. These transtensional basins, however, were apparently separated by areas of transpression, and the Cobequid-Chedabucto fault zone itself was apparently a locus of both transpression and transtension, marked by alternating restraining and releasing bends (Webster et al. 1998; Reynolds et al. 2004; Waldron et al. 2005).

8 The Middle Carboniferous (middle Visean) Windsor Group unconformably to disconformably overlies the Horton Group in much of the Maritimes Basin. It formed when the restricted and hypersaline Windsor Sea flooded the Maritimes rift valleys from the east. The lower part of the Windsor Group is dominated by anhydrite, salt, gypsum, dolostone, and limestone, whereas the upper part consists of red siltstone and fine-grained sandstone with intercalated marine limestone, dolostone, and gypsum. Evaporites of the Windsor Group form diapiric structures in many areas (e.g., Boehner 1992; Durling et al. 1995; Calder 1998). The occurrence of an extensional detachment fault of possible basin-wide extent within the Windsor Group also has been suggested (the Ainslie Detachment of Lynch and Giles 1995). Following withdrawal of the Windsor Sea, a semi-arid climate persisted through the late Visean when gray and red shale and siltstone, thin sandstone intercalations, and thin oolitic limestone of the Mabou Group were deposited in lakes. A profound change in basin evolution and paleoclimate occurred in the mid Namurian between deposition of the Mabou Group and the disconformably to unconformably overlying Cumberland Group (Calder 1998). Dextral motion between the Avalon and Meguma terranes along the Cobequid-Chedabucto fault system is inferred to have caused basin inversion and deformed to varying degrees the older basin fill (Calder 1998). Following this event, the dominant depositional systems in the Maritimes Basin changed from lacustrine to fluvial, represented by sandstone and coal of the Upper Carboniferous Cumberland Group. In the Late Carboniferous to Early Permian, continental red beds of the Pictou Group were deposited widely across the Maritimes Basin during a period of regional thermal sag and increasing aridity. The final assembly of Pangaea brought regional basin inversion, and a pronounced unconformity is present between Upper Carboniferous–Permian and Triassic strata of the Mesozoic Fundy rift basin.

GEOLOGY OF Isle MADAME AREA

EXTENSIONAL MODEL

27 Certain aspects of these relations fit an extensional model in which the conglomerate units are syn-extensional deposits:

• 1) The synchronous development of faults and conglomerate units, as evidenced by clasts of fault rock in conglomerate.
• 2) The juxtaposition of brittle and ductile deformation structures and of rocks of widely different metamorphic grades, suggesting juxtaposition of different crustal levels.
• 3) The systematic presence of thorough cataclasis of the older, deeper assemblage, particularly on its faulted boundaries.
• 4) The high angle between bedding in conglomerate and the cross-cutting faulted boundaries.
• 5) The thickness of conglomeratic debris.
• 6)The dip-slip motion on many fault boundaries.

Display large image of Figure 9

28 In an extensional model, the elements fit together (Fig. 11) via listric detachment faults, with brittle-ductile transitions on the footwall from chloritic breccia to semi-ductile mylonite, fanning dips in syn-extensional upper-plate conglomerate units, domino-style upper-plate faults, and consistent angles of 50–70 degrees between detachments and bedding in conglomerate units (Seranne and Seguret 1987; Howard and John 1987; Lister and Davis 1989; Spencer and Reynolds 1989, 1991; Dickinson 1991). These angles preserve the original intersection of depo-sitional bedding with fault scarps during transport along listric faults (i.e. the contacts originated as buttress unconformities). At the surface, relations and sedimentary patterns are those of half-grabens (see Leeder and Gawthorpe 1987).

29 However, some elements on Isle Madame differ from the model:

• 7) The dips of most of the faults are steep, unlike the gently dipping detachments described in the references cited above.
• 8) Considerable thicknesses of conglomerate-dominated successions are locally overturned

30 We think the simplest explanation of these anomalies is that the expected extensional-basin relations have been further disturbed in some areas by a second stage of extension. In the areas with only one stage of extension, dips of the conglomerate-dominated successions are to the north, and detachment faults would be gently dipping (if they were exposed). The implied detachment headwall is outside the study area, north of Isle Madame, and tectonic transport was north to south. The northward dip, probably gentle, of the detachment in the northern part of the study area suggests local footwall uplift. In areas that experienced a second stage of extension, relations were rotated en masse (as shown in the inset of Fig. 11), causing older detachment faults to dip steeply north, and bedding in the Horton–lower Windsor Group succession to become overturned (but still younging north). Only the younger conglomerate-dominated succession within the second-stage allochthon dips north at moderate angles. Such overturning relations in extension have been observed elsewhere, especially where multi-stage (Hagstrum et al. 1987; Cox and Force, in press). We have considered the possibility that this second-stage event is not extensional but related to inversion; however, that model is not consistent with either the sediment provenance or dip angle of the younger succession. Indeed, it seems more likely that the second stage was related to secondary breakaway in a single extensional episode (e.g., Spencer and Reynolds 1989).

TECTONIC INTERPRETATION FOR Isle MADAME

31 Figure 12 shows an interpretive N–S cross-section of Isle Madame. Where a younger detachment episode has up-ended and repeated an older detachment system, the older detachment fault is exposed as the fault rocks on the northern edge of basement in each domino-style fault block of the younger system. This basement is, of course, the former lower plate of the older detachment system. The older syn-extensional deposits of Horton and lower Windsor groups have been overturned where they have been rotated more than a total of 90° in the two stages. The younger upper Windsor and Mabou Group beds that formed during second-stage extension are not overturned, and outside the younger second-stage detachment allochthon, where the older beds have been tilted only once, they are not overturned either. All of the successions remain north younging.

Display large image of Figure 10

Display large image of Figure 11

32 The second-stage fault cutting the older detachment and basement, although unusual, is not unexpected or unique. Secondary breakaways by definition cut older detachments and may cut basement (Spencer and Reynolds 1989). Lister and Davis (1989) described single-stage fault-locus migrations that include younger faults stranding basement and older detachments on their upper plates.

33 The Crid Islands–Irish Point area (Fig. 2) presents an illuminating glimpse of the extensional relations we propose (Fig. 13). The Shaw Point–Irish Point and Arichat basement belts are both present, separated by two conglomerate-dominated successions. The older Horton and lower Windsor succession (Fig. 6) is vertical to slightly overturned, separated from younger gently dipping upper Windsor and Mabou succession by a younger upper-plate fault with gypsiferous gouge and an inferred unconformity. The older detachment is well exposed on the Crid Islands (Fig. 8) and the younger detachment at Irish Point.

34 This interpretation of extensional history divides Windsor Group deposition into two different tectonic settings, one during first-stage extension and one during the second stage. First stage extension was apparently Tournasian–Visean in age, and the second was Visean–Namurian. The expected unconformities between the two tectono-stratigraphic packages are poorly represented, because they largely occupy different fault blocks and/or different coastal sections. However, the block north of the Arichat basement belt seems to contain an unconformity between Horton and upper Windsor Group rocks (Figs. 13, 14).

35 In this view of the tectonic history of Isle Madame, the only basement block that is “rooted” is that between Shaw and Irish points. Aeromagnetic maps of the area (Geological Survey of Canada 1964) are consistent with this interpretation in showing that only this basement block has any magnetic expression.

36 Our interpretation of multi-stage extension is illustrated on a tectonic map of southeastern Isle Madame showing tilt domains (Fig. 14). The rocks that belong to an older tilt domain, where undisturbed by later extension, are characterized by north dips that average about 50° in syn-extension deposits. In the younger Cap Auguet extensional allochthon, however, further movement has rotated these older deposits so that they are vertical to overturned. Deposits synchronous with younger extension in this area also dip north an average of about 50°. Reconstruction of the second-stage extension in cross-section (Fig. 15) suggests that relations similar to those of meta-morphic core complexes may have existed in the area (e.g., Davis 1987; Lister and Davis 1989), but were fragmented by second-stage faulting. The gentle northward dip of the first-stage detachment suggests local footwall uplift. However, the scarcity of detachment-related ductile mylonite requires that uplift was not from great depth (e.g., Davis 1987; Davis et al. 2004). Second-stage extension itself may be a secondary breakaway as commonly observed where the footwall is uplifted (Spencer and Reynolds 1989).

Display large image of Figure 12

DISCUSSION

37 Extensional basins of Horton Group age are reported elsewhere in the Maritime Provinces as described above, mostly as half-graben fills. Detachments faults have been mapped on Cape Breton Island north of the study area by Lynch (2001); his Ainslie detachment fault is younger and probably related to décollement along evaporite horizons, but his Margaree detachment could be coeval with the older detachment that we describe. The extent of the Margaree detachment system has not been documented to include southern Cape Breton Island, however, and correlation is as yet unwarranted.

38 The geologic setting of Nova Scotia during deposition of the Horton Group involved dextral movement of incoming Meguma terrane against already-docked Avalon terrane (Fig. 1). The association of transtensional basins with transcurrent faults, even in transpressional settings, is widely reported elsewhere, associated with releasing bends of bounding faults (Bradley 1982; Mann et al. 1983) and with unravelling escape structures (Mann 1997). The basins reportedly include pull-apart basins, transform-normal extensional basins, and opening escape triangles.

Display large image of Figure 13

Display large image of Figure 14

39 Within transtensional basins, faults commonly show dip-slip movement (Aksu et al. 2000), especially on structures oblique to master bounding faults. The latter clearly have an older history related to transcurrent movement, but as basin-bounding faults they acquire a dip-slip component as an overprint. They also become listric and semi-ductile at depth as reported by Aksu et al. (2000). Listric faults generate dips in subsiding basin fill. Sub-basins may be separated by narrow horsts outlined by faults of the negative flower structure type. The resulting geometry is similar to that of purely extensional basins, even to the predominance of dip-slip movement on internal faults. Indeed, much of the literature on extensional tectonics has actually come from transtensional settings (reviewed by Dickinson 2002). Where master bounding faults with transcurrent histories are not exposed, it may be difficult to discern a difference between purely extensional and trans-tensional basins. No such faults are obviously exposed in the Isle Madame area, although the Lennox Passage Fault (Fig. 2) is a candidate. However, the fault rocks described here that are interpreted to represent listric detachments of the first stage of extension may themselves represent the dissected remnants of original bounding faults no longer exposed as continuous features. This possibility is consistent with horizontal and oblique slickenlines locally observed in the fault rocks (in contrast to dip-slip slickenlines predominant in the basin fill). If a pull-apart basin is the appropriate model, the fault rocks presumably represent its northern margin because tectonic transport was north to south.

40 Perhaps the complex history of Isle Madame extension is fortuitous; if extension had not been multi-stage we would have been unable to see up-ended deeper elements of a transtensional basin. The next challenge as we see it is to use details of the Isle Madame geology as a tool to elucidate larger regional questions. For example, elsewhere in the Maritimes, transtensional basins are of the same age, but have been described as half-grabens based on shallow-level exposures. Can Isle Madame inform us about the deeper roots of some of these basins? And if so, do these integrated basin geometries have implications for the kinematics of basin formation?

Display large image of Figure 15

CONCLUSIONS

41 The thick conglomerate-dominated successions of Isle Madame are the sedimentary consequences of two periods of lower Carboniferous extension. Deposition was apparently in transtensional basins formed along the Meguma-Avalon ter-rane boundary. The appreciable dips of the sedimentary succes sions result from downfaulting on listric faults. The long narrow belts of basement rocks and associated fault rocks record the juxtaposition of different crustal levels along detachment faults formed during the first period of extension and up-ended by the second. The two-stage extension has exposed relations that may be typical of deeper levels of half-grabens of the Maritimes Basin.