Detrital zircon ages from the lower part of the Late Proterozoic(?) to Middle Cambrian Goldenville Group in the Meguma terrane of Nova Scotia suggest derivation from local sources in the Avalonian and Pan-African orogens on the margins of Early Cambrian Gondwana. Samples from near the top of the group show a broader distribution, including ages back to Archean. The εNd data show a corresponding trend, from slightly positive in the lower Goldenville Group to highly negative in the upper Goldenville Group and overlying Upper Cambrian to Lower Ordovician Halifax Group. The trends are consistent with deposition of the lower part of the Meguma succession in a rift, in which uplifted rift-flanks were the main source of the early basin fill, whereas subsequent thermal subsidence of rift margins allowed for more widespread sediment sourcing in younger units. The rift was possibly located between Gondwana and Avalonia, and may have been the locus for separation of Avalonia from Gondwana to form part of the Rheic Ocean.
Meguma, the most outboard terrane of the Canadian Appalachians (Fig. 1a), has no clear correlative elsewhere in the Appalachian–Caledonide orogen. Although generally regarded as a peri-Gondwanan terrane, Meguma shows marked contrasts with adjacent Avalonia, and its origin has been controversial. Some authors (e.g., Schenk 1997 and references therein) have proposed that Meguma originated adjacent to northwest Africa and was transferred to Laurentia during the Acadian orogeny, whereas others have proposed an origin juxtaposed with western Amazonia (e.g., Keppie 1977) and (or) that Meguma travelled with Avalonia during the early Paleozoic (Murphy et al. 2004a).
We report here the results of U–Pb dating of detrital zircon and Sm–Nd isotopic analyses from Cambrian–Ordovician metasedimentary units in Meguma that are relevant to this controversy. The data demonstrate that sediment provenance changed rapidly from a restricted and juvenile source to more diverse and isotopically evolved sources with abundant Paleoproterozoic zircon, consistent with formation on the Gondwanan margin in an evolving rift that subsequently became inactive and underwent thermal subsidence.
Meguma is characterized by unique stratigraphy, including a thick (>10 km) Cambrian (and possibly older) to Early Ordovician turbiditic clastic succession, historically assigned to the Meguma Group, and divided into a lower, coarser grained Goldenville Formation and an upper, dominantly fine-grained Halifax Formation. Stratigraphic work (O’Brien 1988; Waldron 1992) and recent mapping (e.g., White 2007) resulted in subdivision of these thick formations and their elevation to group status. The entire package now is termed the Meguma Supergroup (White 2008).
The Goldenville Group is dominated by psammite (metasandstone), with subordinate pelitic rocks (metasiltstone, slate, argillite). A locality close to the top of the group (Figs. 1a, 1b) yielded a Middle Cambrian trilobite faunule (Pratt and Waldron 1991). Nearer the base of the exposed Goldenville Group (Fig. 1b), the trace fossil Oldhamia indicates an Early Cambrian or possibly late Neoproterozoic age (White et al. 2005). The overlying Halifax Group is dominated by pelite, associated with fine- and very fine-grained metasandstone. Rare graptolites high in the Halifax Group (Fig. 1b) indicate an Early Ordovician age (e.g., Cumming 1985).
Northwest of the Chebogue Point shear zone (CPSZ; Fig. 1a), the Halifax Group is overlain unconformably by Silurian volcanic and sedimentary rocks assigned to the White Rock Formation and equivalent units, deposited in a shallow-marine rift environment (MacDonald et al. 2002, and references therein). The overlying Early Devonian Torbrook Formation records shelf conditions (Jensen 1975). In addition to the contrast in stratigraphy (Fig. 1b), this part of Meguma is characterized by abundant mafic sills (White and Barr 2004; White et al. 2006). Furthermore, our limited paleocurrent data from the area indicate southward paleoflow (Fig. 1a), in contrast with northwestward to northeastward flow reported elsewhere in the terrane (Schenk 1970; Waldron and Jensen 1985; Waldron 1988).
The basement of Meguma is poorly known. Sm–Nd data from mafic and felsic volcanic rocks in the White Rock Formation are mainly positive, suggesting crustal contamination of mainly mantle-derived magmas (MacDonald et al. 2002) but also consistent with sources in Avalonian basement (Keppie et al. 1997). Greenough et al. (1999) recorded concordant and slightly discordant zircons with Avalonian and Pan-African (630–575 Ma) U–Pb ages in xenoliths thought to be derived from basement; they also inferred Mesoproterozoic components in basement from upper intercept ages projected through discordant zircon fractions.
Sample locations and methods
Four metasandstone samples for detrital zircon dating were collected in the Goldenville Group (Fig. 1a, 1b). Locations were chosen to cover a wide age span, to include dated fossil localities, and to complement the previous small data set of Krogh and Keppie (1990). Zircon grains were extracted, imaged by electron backscatter, and dated by laser ablation multicollector–inductively coupled plasma–mass spectrometry (MC–ICP–MS) using the methods of Simonetti et al. (2005). Results are shown as probability density plots in Fig. 2, excluding grains with >10% discordance. These samples were also analysed for their Sm–Nd isotopic composition, together with additional samples to extend coverage across the CPSZ, and into the overlying Halifax Group. These data are combined with published Sm–Nd data, all from east of the CPSZ, but not previously positioned stratigraphically. Summary results are shown in Figs. 1c and 2; full tables of results appear in the supplementary data.
U–Pb detrital zircon dating
Sample 1, from the lowest point in the exposed stratigraphy west of the CPSZ (Fig. 1b), shows a strongly clustered age distribution; 71 of 76 grains lie between 750 and 540 Ma. The density distribution has a strong peak at ∼640 Ma, with subsidiary peaks at 560, 585, and possibly 700 Ma. The youngest grain, at 544 ± 18 Ma, represents the maximum depositional age of the sample and hence of the exposed Goldenville Group. Of the remaining grains, two are Neoproterozoic, and three are Paleoproterozoic.
Sample 2, from a metasandstone bed ∼4 km higher in the section, an Oldhamia-bearing interval of fine-grained metasedimentary rocks known as the High Head member (White et al. 2005), shows an even more restricted detrital age range; all 124 grains give ages between 740 and 540 Ma, with distinct peaks at 560, 580, and 620 Ma. A grain at 537 ± 15 Ma provides the most reliable maximum depositional age.
Sample 5, from the stratigraphically highest metasandstone unit in the Goldenville Group northwest of the CPSZ (Figs. 1a, 1b), shows a much more diverse age spectrum. A large Neoproterozoic cluster (78 of 125 grains) has peaks at ∼550, 590, and possibly 620 Ma. A concordant grain at 529 ± 19 Ma represents the best constraint on depositional age. A second large peak occurs at 2060 Ma. Scattered grains show Mesoproterozoic and older Paleoproterozoic ages. Seven Archean grains range from 2620 to 3036 Ma.
Sample 13 is from an equivalent stratigraphic level east of the CPSZ (Fig. 1b), adjacent to a fossiliferous Middle Cambrian bed (Pratt and Waldron 1991). Of 47 grains with <10% discordance, 12 are late Neoproterozoic, grouped at ∼560, 585, and 620–640 Ma. The remaining 35 grains range from older Neoproterozoic to Mesoarchean. Abundant Paleoproterozoic grains include six grains overlapping within error at ∼2135 Ma. Seven Archean grains show ages scattered from 2540 ± 17 to 3110 Ma.
Sm–Nd isotopic analyses
The εNd data (calculated at 540 Ma) west of the CPSZ show a clear change from slightly positive (+0.78 and +0.86, indicating relatively juvenile sources) in lower samples, to highly negative values (–6.43 to –8.79, consistent with more evolved sources) at stratigraphically higher levels. The εNd values are strongly correlated with detrital zircon ages; the two samples (1 and 2) with positive values yielded highly restricted, young detrital zircon populations (Fig. 2). An upward decreasing trend continues into the Halifax Group. More data are available east of the CPSZ, where the upward trend toward negative εNd is also clear, reaching highly negative values (∼–16).
As expected, Meguma detrital zircon age distributions do not resemble those from the Cambrian of Laurentia, which are characterized by strong peaks at 1.0–1.3 Ga, and which typically lack zircons from 2.0 to 2.5 Ga (e.g., Cawood et al. 2007, and references therein). Our older samples (1 and 2) show restricted, dominantly late Neoproterozoic age distributions that resemble some Cambrian units from Avalonia; for example, a sample from Great Britain shows peaks at 550, 560, 605, and 620 Ma (Murphy et al. 2004b). These distributions suggest derivation from local sources in the Avalonian – Pan-African orogen. Younger samples (5 and 13) show lower εNd and a broader distribution of detrital zircon ages. These data somewhat resemble results of Krogh and Keppie (1990) (Fig. 2). However, both of our samples 5 and 13 contain rare Mesoproterozoic grains not recorded in the Krogh and Keppie (1990) data. Five grains in sample 5, and three more in sample 13, span a range from 1150 to 1481 Ma. In addition, sample 13 contains three grains with typical “Grenville” ages between 1000 and 1050 Ma. Younger sedimentary units in Meguma are similar (Fig. 2): the overlying White Rock and Torbrook formations studied by Murphy et al. (2004a) have abundant zircon grains between 540 and 650 Ma, and scattered Mesoproterozoic (1000–1500 Ma) and Paleoproterozoic (1900–2200 Ma) grains. Murphy et al. (2004a) argued that the Mesoproterozoic grains in these units require an Avalonian source and suggested on this basis that Meguma and Avalonia were contiguous in the mid-Paleozoic. Our data show that the mid-Paleozoic units contain distributions closely similar to those in the underlying Meguma Supergroup; their zircon grains could easily have been derived by erosion within Meguma, potentially at the unconformity above the Meguma Supergroup, and hence do not require a contiguous Avalonian source.
Sources for 2000–2200 Ma zircon are rare in Laurentia but common in parts of Gondwana, such as the Eburnian and Birimian orogens of West Africa (Rocci et al. 1991; Lerouge et al. 2006). Similar age sources are recorded from the Congo craton (e.g., De Waele and Fitzsimons 2007), though associated with a large ∼1800 Ma peak. In South America, ages of 2000–2200 Ma are found in the Amazonian (e.g., Neves et al. 2006) and Sao Francisco (Barbarosa and Sabaté 2002) cratons. Detrital zircons from the La Cebila formation of Argentina (Finney et al. 2003, 2004) also show a striking peak in this range, as does the Gondwanan Suwanee terrane of Florida (Mueller et al. 1994). Age distributions from European Armorica also show a strong 2.0–2.2 Ga peak (Linnemann et al. 2004).
Nance and Murphy (1996) and Linnemann et al. (2004) suggested that abundance of Mesoproterozoic zircon may be used to differentiate peri-Gondwanan terranes originating adjacent to Amazonia from those originally adjacent to West Africa. The abundance of Eburnian and relative scarcity of Grenvillian detrital zircon in our Meguma samples suggest an origin adjacent to West Africa, but within reach of occasional Amazonian grains. An original position adjacent to Morocco (e.g., Schenk 1971) seems unlikely because Moroccan Cambrian successions are dominated by warm-water carbonates, in contrast to clastics and cool-water carbonates in Avalonia and Meguma (e.g., Landing 1996; Linnemann et al. 2004). An origin on the north African margin is conceivable, as 1 Ga zircon is abundant in Cambrian–Ordovician sandstone from Israel and Jordan (Kolodner et al. 2006), but the Meguma distributions do not resemble those reported from central European terranes believed to have originated along this part of the Gondwanan margin (Fig. 3).
The upward trend in our data, from relatively juvenile Pan-African and Avalonian sources to more diverse, older populations is mirrored in the εNd data, indicating that it is a general feature of the Meguma Supergroup. The trend is most easily explained by deposition in a rift or other extensional environment that subsequently became inactive; uplifted flanks were the main early source, whereas subsequent thermal subsidence allowed more widespread sourcing. A rift environment helps to explain the thickness (>10 km) and rapid accumulation (>250 mm/ka) of the Meguma Supergroup. Mafic intrusions west of the CPSZ (White and Barr 2004), and stratigraphic contrasts across the zone, also are consistent with a rift. Although no coarse or subaerial clastic sedimentary rocks have been described from the Meguma succession, its base and margins are unexposed; we would predict less mature sediments at depth below the current deepest level of exposure, consistent with the trends in εNd data (Fig. 1c). At least one margin of the proposed rift must have provided a cratonic source for the older zircons observed; however, the postulated rift may have developed within a continental margin setting and might have formed in a transtensional environment as envisaged for Cambrian Avalonia (e.g., Nance and Murphy 1996).
Similarities to Avalonian detrital zircon populations, juvenile isotopic compositions, and the consistency of sparse faunal evidence with Avalonian affinity (Pratt and Waldron 1991) favour a location generally between Gondwana and Avalonia. Although our data suggest that the Meguma rift became inactive during the Middle Cambrian, they do not indicate whether this change represents a transition to a passive margin environment or whether it indicates that the later Meguma sediments were deposited in an intracratonic “failed rift.” Subsidence analyses from Avalonia suggest that its separation and the opening of the Rheic Ocean did not begin until the Early Ordovician (e.g., Prigmore et al. 1997), supporting the idea that the Early Cambrian rift was a failed one. However, other workers (Landing 1996, 2005) argue that Avalonia was separated from Gondwana by the Early Cambrian, in which case our data may record the opening of part of the Rheic Ocean. Numerous published reconstructions of peri-Gondwanan terranes in the Early Paleozoic differ in detail, but agree that Avalonia and related Appalachian terranes were located along a previously active margin of Gondwana (e.g., Nance and Murphy 1996; Linnemann et al. 2004; Murphy et al. 2004c, 2006). In Fig. 3, we therefore speculate that Meguma was originally located between Avalonia and West Africa, in the rift system along which the Rheic Ocean opened. Present data are insufficient to determine whether Meguma accompanied Avalonia after the opening of the Rheic Ocean or stayed behind nearer Gondwana.
The lowest exposed units of the Meguma Supergroup of southern Nova Scotia display detrital zircon populations and Sm–Nd isotopic characteristics consistent with derivation from the Avalonian – Pan-African orogen. Upsection, far-travelled ancient detritus from West Africa and (or) Amazonia becomes abundant. Paleoproterozoic grains everywhere outnumber Mesoproterozoic “Grenvillian” grains. These characteristics indicate that during latest Neoproterozoic and Early Cambrian time Meguma was probably located in a rift between West Africa and Avalonia, perhaps the one along which Avalonia separated in the Early Cambrian or Early Ordovician. Age distributions previously reported from Silurian to Devonian units in Meguma are consistent with derivation from younger parts of the Goldenville Group; Meguma, therefore, did not necessarily travel with Avalonia during the mid-Paleozoic.
Financial support was provided by Natural Sciences and Engineering Research Council of Canada grants to JWFW, SMB, and LMH and the Nova Scotia Department of Natural Resources (NSDNR) operating budget for CEW. We appreciate the assistance of Judy Schultz in preparing samples. Damian Nance, journal referees Meg Thompson and Cees van Staal, and Associate Editor Brendan Murphy made suggestions that improved the manuscript.
1 Supplementary data can be found on the journal Web site (http://cjes.nrc.ca) or may be purchased from the Depository of Unpublished Data, Document Delivery, CISTI, National Research Council of Canada, Building M-55, 1200 Montreal Road, Ottawa, ON K1A 0R6, Canada (DUD 3903). For more information on obtaining material, refer to http://cisti-icist.nrc-cnrc.gc.ca/unpub_e.shtml.
- Received October 6, 2008.
- Accepted January 27, 2009.
- Published on the NRC Research Press Web site at http://cjes.nrc.ca on March 2, 2009.