Late Devonian shear-zone reactivation in the Canadian Appalachian orogen Open Access

Supplementary material: All original data are available at https://osf.io/KEDJP/ [last accessed 16 July 2023].

The Late Devonian evolution of the Canadian Appalachians is marked by the oblique collision of the Meguma terrane with the composite Laurentian margin, initiating the Neoacadian Orogeny and the simultaneous opening of the Maritimes Basin (Fig. 1) (Murphy and Keppie 1998). During this collision, the rocks located in the Meguma terrane and in its suture with Avalonia were rapidly exhumed (Archibald et al. 2018) by transpressive deformation along the Cobequid–Chedabucto Shear Zone (Waldron et al. 1989). Similarly, shear-zone reactivation at the composite margin of Laurentia accommodated some rapid exhumation during the Late Devonian (Kellett et al. 2014; Willner et al. 2018) but relatively little work has been done to test the distribution of transpressive and transtensive deformation with respect to segments of different orientations. This inboard reactivation may be key to understanding the tectonic setting of the Neocadian Orogeny in areas where terrane-bounding shear zones are not exposed for field investigations. In addition, fault reactivation may also influence the emplacement of magma inboard of the suture (e.g. Sombini dos Santos et al. 2020) or outboard in the Meguma terrane (e.g. Bickerton 2021) and thereby intrusion-related mineralization (Kellett et al. 2021).

Here we characterize the kinematics and timing of deformation of the Eastern Highlands Shear Zone (EHSZ) and the Coinneach Brook Shear Zone (CBSZ) on Cape Breton Island, Nova Scotia (Fig. 2) in order to determine if these structures were coeval during the Late Devonian and if their combined kinematics may have accommodated rapid Late Devonian exhumation of the Cape Breton Highlands, as recently described by McCarron (2020).

Our study builds on extensive earlier work on the EHSZ (Barr and Raeside 1989; Dunning et al. 1990b; Raeside and Barr 1990; Lin 1993, 1995, 2001; Chen et al. 1995) and on the mapping by Horne (1995) along the CBSZ. We apply a combination of field mapping, outcrop and thin-section structural observations with texturally resolved, in situ U–Pb and ⁴⁰Ar/³⁹Ar petrochronology to characterize the timing, kinematics and conditions of the different episodes of deformation across these shear zones. These datasets provide the means to compare reactivation of the EHSZ and the CBSZ with the development of the terrane-bounding Cobequid–Chedabucto shear zone. The results of this comparison help to determine if the EHSZ and CBSZ were (re)activated during the Neoacadian Orogeny and if they accommodated transpressive or transtensive kinematics. In a broader context, this contribution also helps to characterize the potential of pre-existing, inboard structural weaknesses in accommodating oblique convergence during terrane accretion.

The Appalachians accretionary orogen formed through the Late Cambrian–Carboniferous convergence of Laurentia and Gondwana. From a North American perspective, accretion began as several oceanic arcs that formed near Laurentia, or rifted from its margin, were accreted onto the Laurentian passive margin during three co-temporal, but distinct, phases of the Taconic Orogeny (van Staal et al. 2007). These were followed by the successive collisions of the peri-Gondwanan microcontinents: Ganderia in the Late Ordovician, Avalonia in the Silurian–Early Devonian and Meguma in the Late Devonian, producing the Salinic, Acadian and Neoacadian orogenies, respectively (Dunning et al. 1990a; Hibbard 1994; van Staal et al. 2009). Finally, composite Laurentia collided with a Gondwanan Promontory in the southern Appalachians (USA) during the Carboniferous Alleghanian Orogeny (Secor et al. 1986a, b).

The Acadian suture, situated between Avalonia and the previously accreted Ganderia, is interpreted to correspond to the Caledonia Fault in New Brunswick, the MacIntosh Brook Fault in Cape Breton Island and the Hermitage Bay–Dover Shear Zone (HDBSZ) in Newfoundland (Fig. 1) (van Staal et al. 2009). The timing of Acadian sinistral deformation in the HDBSZ has been bracketed between c. 422 ± 4 Ma, a U–Pb date from zircon rims in a deformed granite dyke, and 392 ± 2 Ma, the ⁴⁰Ar/³⁹Ar dates from white mica in the same rocks (Kellett et al. 2016). Sense and timing of ductile shear along the Caledonia and MacIntosh Brook faults is uncertain and overprinted by Carboniferous brittle deformation (Raeside and Barr 1990; Waldron et al. 2015).

The dextral-transpressive Cobequid–Chedabucto shear zone is the onland expression of the Minas Fault, interpreted as the Neoacadian suture that separates the Meguma and Avalonia terranes (Murphy et al. 2011). Based on the presence of a magmatic foliation and corresponding magnetic anisotropy, the 378.7 ± 1.2–371.8 ± 0.8 Ma South Mountain Batholith (chemical abrasion isotope-dilution thermal ionization mass spectrometry (CA-ID-TIMS) on zircon: Bickerton et al. 2022), located south of the Cobequid–Chedabucto shear zone in the Meguma terrane, is interpreted to have been emplaced during the Neoacadian Orogeny (Horne et al. 1992; Benn et al. 1999). Moreover, biotite from a deformed granodiorite of the West Moose River pluton that intruded the CCSZ during the Late Devonian provides an ⁴⁰Ar/³⁹Ar age of 337 ± 4 Ma, interpreted to represent the last ductile motion along the shear zone (Pe-Piper et al. 2004, 2017; Archibald et al. 2018). The foliated Kelly Brook pluton intruded the CCSZ at c. 375 ± 5 Ma (U–Pb zircon). It yielded overlapping ⁴⁰Ar/³⁹Ar synkinematic muscovite and U–Pb syn- to post-kinematics apatite ages of 369 ± 1 and 361 ± 6 Ma, respectively, indicating rapid exhumation during deformation of the shear zone (Archibald et al. 2018). Similarly, the South Mountain Batholith crystallized at c. 380–370 Ma (U–Pb on zircon: Bickerton et al. 2022) and yielded synchronous ⁴⁰Ar/³⁹Ar muscovite cooling ages of 381 ± 1 and 371 ± 2 Ma (Keppie et al. 1993; Reynolds et al. 2004). These cooling ages indicate rapid exhumation that led to the formation of extensional jogs that are marked by Late Devonian–Early Carboniferous Horton Group sedimentary rocks and by 365 ± 4–358 ± 4 Ma bimodal Cape Chignecto plutons (granite and gabbro: Murphy et al. 2011) (see also Doig et al. 1996; Dunning et al. 2002; Pe-Piper et al. 2004).

Overall, geochronological and kinematic data indicate that the Neoacadian suture zone recorded ductile transpressive deformation between 378.7 ± 0.6 and 337 Ma ± 4 Ma, comprising oblique dextral strike-slip deformation along the CCSZ and uplift of Meguma terrane relative to composite Laurentia. This time period also corresponds to the dextral transpressive reactivation of the HDBSZ at c. 385 ± 4 Ma (Kellett et al. 2016), leading to the rapid exhumation of the stitching Ackley Granite indicated by overlapping U–Pb zircon (377 ± 4 Ma) and ⁴⁰Ar/³⁹Ar total fusion on hornblende (375 ± 4 Ma) and biotite (372 ± 5 Ma: Kontak et al. 1988; Kellett et al. 2014).

The EHSZ is located within the Ganderia terrane on Cape Breton Island in Nova Scotia, Canada (Fig. 1). It marks the contact between the Ordovician–Silurian volcanic and sedimentary rocks and the Devonian granitoids of the Aspy terrane to the NW and the Late Precambrian metasedimentary rocks and late Proterozoic dioritic–granitic plutons of the Bras d'Or terrane to the SE (Fig. 2a, b) (Barr and Raeside 1989). The EHSZ records two episodes of deformation. The first episode of deformation (D₁) is characterized by SE-(Bras D'Or)-side-up kinematics marking the initial juxtaposition of the Aspy and Bras d'Or terranes (Lin 1993). It is followed by the intrusion of the 375 + 5/ − 4 Ma Black Brook Granitic Suite (U–Pb monazite: Dunning et al. 1990a; Lin 1993). The second episode of deformation (D₂) is characterized by an oblique composite deformation with SE-side-down dextral kinematics that reactivated the EHSZ (Fig. 2b) (Lin 1995; Piette-Lauzière et al. 2020). This deformation extended into the Black Brook Granitic Suite to form the Roper Lake Shear Zone (RLSZ) (Fig. 2b) (Lin et al. 1998). While the timing of the first deformation event along the EHSZ is inferred to 425–415 Ma based on amphibole ⁴⁰Ar/³⁹Ar geochronology across the structure (Lin 2001), the age of the second deformation remains unconstrained.

The CBSZ is a north-trending structure hosted in the 430 ± 2 Ma Taylors Barren pluton in the Aspy terrane (U–Pb on zircon: Horne et al. 2003) and into the metasedimentary rocks of the Jumping Brook Metamorphic Suite. It is intruded by the 363 ± 2 Ma Margaree pluton (Fig. 2b) (Sombini dos Santos et al. 2020). This structure is characterized by oblique dextral west-side-up movement (Mengel et al. 1991; Horne 1995) and may have deformed synchronously with the EHSZ (Lynch 1996). However, there is a paucity of information regarding the timing of this deformation and whether its development was coeval with either of the deformation events recorded in the EHSZ.

The SE-side-up kinematics and ⁴⁰Ar/³⁹Ar amphibole geochronology across the EHSZ indicates that the structure contributed to the exhumation of the Bras-d'Or terrane during the Silurian (D₁: Lin 2001). The Late Devonian evolution of the Cape Breton Highlands is also characterized by the rapid exhumation of the Aspy terrane. This exhumation is marked by overlapping dates from the same area with monazite and apatite ages of 389 ± 3 and 379 ± 3 Ma, respectively (U–Pb, apatite discordia from analyses of three specimens: McCarron 2020), and ⁴⁰Ar/³⁹Ar hornblende cooling ages ranging from 393 ± 3 to 355 ± 4 Ma (Price et al. 1999) that overlap with ⁴⁰Ar/³⁹Ar biotite cooling ages ranging from 388 ± 3 to 371 ± 3 Ma (Reynolds et al. 1989).

We sampled a variety of rock types within the EHSZ and in the CBSZ to characterize and date the regional deformation, cooling and hydrothermal alteration histories. U–Pb zircon geochronology was used to determine the crystallization age of igneous rocks and dykes that have various cross-cutting relationships with the regional and shear-zone fabrics. We also performed in situ U–Pb geochronology on monazite, xenotime and apatite associated with the main shear-zone foliation and locally developed hydrothermal assemblages. The laboratory analyses were performed over seven analytical sessions with four different instrumentation set-ups. A specimen list, along with the analytical session, instrumentation and reference material information for U–Pb geochronology, is provided in Table 1. Thin-section backscattered electrons (BSE) images showing grain locations, high-contrast grain images, elemental maps and analytical data are provided in the Supplementary material. During analysis, we monitored the accuracy and precision of our results using secondary reference materials of known ages. Their measured and accepted ages are also reported in Table 1. Specimen descriptions are provided in the Results to contextualize the individual ages. In addition, we investigated the deformation and/or cooling age of the shear zones using muscovite in situ⁴⁰Ar/³⁹Ar thermochronology on thick sections of mylonitic specimens and step-heating experiments on single grains of hornblende, muscovite and biotite. Data tables of these analyses are provided in the Supplementary material. Complete method descriptions are provided in  Appendix A.

We analysed 12 specimens from the EHSZ and the CBSZ using U–Pb zircon, monazite and xenotime ages, and apatite geochronology, as well as ⁴⁰Ar/³⁹Ar hornblende, muscovite and biotite geochronology, to determine the maximum and minimum ages of D₁, D₂, cooling and hydrothermal alteration. Here we present these results with respect to the structural context of the samples and their relationships with the regional deformation of the Cape Breton Highlands, then D₁, D₂ and hydrothermal alteration of the EHSZ, and with the deformation of the CBSZ.

Specimen 16-30A (Fig. 3a) was sampled from a quartz-diorite orthogneiss located 5 km NW of the EHSZ (Fig. 2b). The orthogneiss is cut by a 10–15 cm-wide dyke of foliated quartz-diorite orientated parallel to the main outcrop foliation (specimen B in Fig. 3a). That foliated dyke, in turn, contains non-foliated pegmatite in its central portion (specimen C in Fig. 3a). In thin section, specimen A displays a lepidoblastic foliation defined by aligned biotite grains. Quartz in the specimen is characterized by internal subgrain development, bulging and pinning textures at the grain boundaries. As in specimen A, biotite grains in specimen B have a preferred orientation parallel to the main foliation of the country rock. Quartz grains in specimen B locally contain internal subgrains that form a weak chessboard extinction pattern (Fig. 4a). Finally, specimen C has no obvious foliation, although quartz grains also locally display a chessboard extinction pattern (Fig. 4b). The variable deformation recorded in these specimens make them excellent candidates to quantify the timing of regional deformation. Because of its fabric, the crystallization age of specimen A is expected to inform the maximum age of regional deformation, while the age of the least deformed specimen C is expected to provide information about the minimum age of regional deformation.

Analytical results for 28 of the 40 zircon cores from specimen 16-30A form a cluster with a ²⁰⁷Pb-corrected ²³⁸U/²⁰⁶Pb mean age of 430 ± 2 Ma with a mean square weighted deviation (MSWD) of 1.6 (Fig. 5a), while a younger cluster of five analyses yielded a ²⁰⁷Pb-corrected ²³⁸U/²⁰⁶Pb weighted mean age of 409 ± 4 Ma with a MSWD of 2.1. In specimen 16-30B, the 12 youngest analyses defined a ²⁰⁷Pb-corrected ²³⁸U/²⁰⁶Pb weighted mean age of 382 ± 2 Ma (MSWD = 2.1: Fig. 5b). The four youngest analyses from specimen 16-30C form a cluster that yielded a weighted mean age of 357 ± 7 Ma with an under-dispersed MSWD of 0.16, while the 29 other analyses form a poorly defined cluster with a weighted mean age of 412 ± 2 Ma and a high MSWD of 16 (Fig. 5c). The U and rare earth elements (REE) compositions of the zircon analyses of 16-30A, 16-30B and 16-30C plot along a trend of progressive enrichment in U, REE, and light REE (LREE) with respect to heavy REE (HREE) (Fig. 6).

Specimen 16-27 was sampled from a rhyolite within the EHSZ (Fig. 1b). Because of its fine grain size, no kinematic indicators were apparent in outcrop (Fig. 3b). In thin section, however, the specimen displays C–S–C′ fabrics (Fig. 4c) interpreted by Piette-Lauzière et al. (2019) to indicate NE-side-up kinematics related to D₁ deformation. The crystallization age of this specimen thus provides additional information about the maximum age of D₁. The zircon grains from specimen 16-27 are idiomorphic with oscillatory zoning and minor fractures in grain rims or along their long axes (Fig. 7a). A total of 54 grain cores were analysed in specimen 16-27, from which one severely discordant analysis and five inherited analyses (>550 Ma) were excluded (see the Supplementary material). The 46 remaining analyses yielded a ²⁰⁴Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 424 ± 1 Ma with a MSWD of 5.1 (Fig. 5d).

Specimen 17-38 was sampled from the hinge of a decametre-sized chevron fold marked by the structural contact between an orthogneiss and a metaconglomerate (Fig. 3c). In thin section, the quartz-diorite specimen presents a lineation defined by the long axis of amphibole grains (Fig. 4d). Two hornblende grains were selected for ⁴⁰Ar/³⁹Ar single-grain step-heating analyses until total fusion. The heating schedule was designed to extract the extraneous argon first and then release the radiogenic argon (³⁹Ar). The first grain released 77% of ³⁹Ar over the seventh and eight steps, yielding a pseudo-plateau age of 404 ± 1 Ma (MSWD = 7.73: Fig. 8a); the second grains did not produce a plateau age but yielded a similar isochron age of 403 ± 3 Ma (MSWD = 5.8: Fig. 8b). Specimen 17-107 was collected from the least strained outcrop of the quartz-diorite unit that shows a weak foliation. An amphibole grain released 82% of ³⁹Ar over the last seven heating steps and yielded a similar plateau age of 402 ± 3 Ma (MSWD = 1.9: Fig. 8c). Plateau ages were defined by a minimum of three consecutive steps with overlapping ages at 2 SE that represent at least 50% of the ³⁹Ar released. Heating steps that released less than 0.5% ³⁹Ar were omitted.

Specimen 17-58 was sampled in the EHSZ from a foliated granite dyke with a steep mineral lineation within a deformed psammite (Fig. 3d). In thin section, the specimen presents C–S fabrics consistent with D₂ SE-side-down kinematics (Fig. 4e) (Piette-Lauzière et al. 2019). The zircon grains from this specimen have generally sub-idiomorphic habits with oscillatory zoning visible in BSE images and cathodoluminescence (Fig. 7b). A total of 22 grains were analysed using a sensitive high-resolution ion microprobe (SHRIMP-II), including eight highly radiogenic analyses (U >2000 ppm), which were rejected to avoid potential matrix effects (e.g. Larson et al. 2010), and one inherited analysis. Ten of the remaining 13 grain core analyses defined a ²⁰⁴Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 370 ± 1 Ma (MSWD = 2.5: Fig. 9a).

Six monazite grains between 20 and 50 µm in diameter were identified for in situ geochronology. They are typically xenomorphic in shape and commonly surrounded by an apatite corona (Fig. 7c). Four spot analyses in these grains defined a lower intercept age of 355 ± 6 Ma (MSWD = 1.6) in Tera–Wasserburg space (Fig. 9b).

In addition to the U–Th–Pb geochronology on this specimen, three large (5 mm in length) muscovite mica fish were selected for in situ⁴⁰Ar/³⁹Ar geochronology (Fig. 4e). For 16 of the 19 spot analyses, more than 90% of ⁴⁰Ar released was radiogenic and defined a weighted mean age of 374 ± 1 Ma (2 SE) with a MSWD of 3 (Fig. 10a).

Specimen 17-76 was sampled from a garnet-bearing psammite (Fig. 4f) on the NW flank of the EHSZ (Fig. 2b). Five xenotime grains with diameters ranging from 25 to 100 µm were identified in thin section for in situ geochronology; they are generally xenomorphic. These grains are in contact with biotite and muscovite with a long axis orientated parallel with the foliation (Fig. 7d). Microprobe wavelength dispersive elemental maps show faint zoning in Th and Y (Fig. 7d). Twenty-three spot analyses across the five grains yielded a ²⁰⁷Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 377 ± 2 Ma (MSWD = 1: Fig. 9c).

Specimen 16-44A was sampled from a muscovite-rich pelitic schist in contact with a deformed mafic schist in the EHSZ (Fig. 2b). The contact between the two lithologies is characterized by C–S–C′ fabrics in the pelite (Fig. 11a) indicative of SE-side-down D₂ kinematics (Piette-Lauzière et al. 2019, 2020). In thin section, the pelite is characterized by similar C–S–C′ fabrics and small chevron folds marked by muscovite (Fig. 12a). Eleven monazite and seven xenotime grains, both contained within the muscovite-defined foliation of the specimen, were analysed in situ for geochronology. The monazite grains have a xenomorphic shape with a rough diameter of about 40 µm (Fig. 7e). Microprobe wavelength dispersive spectroscopy elemental maps show non-systematic chemical zoning within the monazite. Thirty-nine spots were analysed across the 11 monazite grains. Three grains were discarded because of their anomalously young date (<100 Ma) and another analysis discarded because it yielded an inherited age (456 Ma: see the Supplementary material). Three of the remaining analyses yielded a young, under-dispersed, ²⁰⁷Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 345 ± 14 Ma (MSWD = 0.0067) and 28 older analyses yielded a weighted mean age of 376 ± 2 Ma (MSWD = 1, n = 28: Fig. 13a). Xenotime grains in the specimen also have a xenomorphic habit, with some displaying a sigma-type shape (Fig. 7f). Eight xenotime grains, with long axes of between 40 and 100 µm, were selected for in situ geochronology. No zoning is apparent in electron microprobe elemental maps (Fig. 7f). A total of 48 spots were analysed in the eight xenotime grains, of which three were discarded because of irregular analytical signals. Forty-two of these analyses defined a ²⁰⁷Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 377 ± 3 Ma (MSWD = 2.8: Fig. 13b), which is identical, within uncertainty, to the monazite age obtained from the same specimen during the same analytical session.

Muscovite grains from the hinge of a kink band, the main foliation plane and a C′ shear band in this specimen were analysed for in situ⁴⁰Ar/³⁹Ar geochronology (Fig. 12a) and yielded similar ages. For nine of the 12 ablated spots, more than 90% of ⁴⁰Ar released was radiogenic and yielded a weighted mean age of 371 ± 2 Ma across all textural positions (MSWD = 1: Fig. 10b). An aliquot of muscovite was also separated for a step-heating experiment and yielded a plateau age of 371 ± 2Ma (MSWD = 1.62) defined by 98% of ⁴⁰Ar released over the last four steps (Fig. 8d). The in situ and plateau ages are identical, and are also within uncertainty of the monazite and xenotime ages calculated for this specimen.

Specimen 17-56 was collected from an outcrop of metaconglomerate within the EHSZ (Fig. 2b). In thin section, it is characterized by a foliation defined by muscovite and biotite intercalated with fine-grained plagioclase and quartz grains (Fig. 12b). The micas also define a C–S–C′ fabric consistent with SE-side-down D₂ kinematics (Piette-Lauzière et al. 2019, 2020). Two medium-sized (0.5 mm-wide) muscovite grains within a C′ plane and one medium-sized muscovite grain within the foliation plane were selected for in situ⁴⁰Ar/³⁹Ar geochronology. A total of 10 spot analyses distributed across the grains yielded an ⁴⁰Ar/³⁹Ar weighted mean age of 371 ± 2 Ma (MSWD = 1: Fig. 10c). An aliquot of muscovite was separated for a step-heating experiment. Of the 14 steps, 93% of the radiogenic argon was liberated over steps 8 and 12–14 and yielded a plateau age of 371 ± 2 Ma (MSWD = 1.6: Fig. 8e), which is identical to the in situ age. An aliquot of biotite was also separated for a step-heating experiment. Similarly, 95% of the radiogenic argon was liberated over the last three heating steps and yielded a plateau age of 375 ± 2 Ma (MSWD = 0.8: Fig. 8f).

Specimen 17-51 was collected from a foliated granite on the Aspy flank of the EHSZ (Fig. 2b). In thin section, the foliation is defined by isolated muscovite and biotite grains (Fig. 12d). Two single-grain aliquots of muscovite and one-single grain aliquot of biotite were analysed for ⁴⁰Ar/³⁹Ar geochronology. The first muscovite step-heating experiment yielded a plateau age of 372 ± 3 Ma over heating steps 4, 5, 7 and 11, representing 60% of the radiogenic argon released (MSWD = 2.8: Fig. 8g). The second muscovite step-heating experiment yielded a plateau age of 370 ± 2 Ma with 87% of the radiogenic argon released over steps 5–10 (MSWD = 1.6: Fig. 8h). For the biotite step-heating experiment, the last seven heating steps yielded a plateau age of 376 ± 2 Ma, representing 64% of the liberated radiogenic argon (MSWD = 1.88: Fig. 8i).

The SE part of the EHSZ is characterized by a thick unit of variably deformed, brecciated and altered granite or rhyolite (Fig. 1b) (Piette-Lauzière et al. 2018). Two specimens were collected near the contact of these rocks with the Cameron Brook Granodiorite to determine the timing of crystallization of this unit, formation of the breccia and alteration. Specimen 17-99B was collected from an outcrop of deformed granite with well-defined D₂ C–S–C′ fabrics (Fig. 11b), while specimen 17-98 was collected 50 m away on the same outcrop near the contact (Fig. 11c, d). In thin section, specimen 17-99B contains brittle–ductile fabrics (Fig. 12c) indicative of dextral SE-side-down oblique kinematics identical to the D₂ kinematics within the ductile portion of the shear zone. The C–S–C′ assemblage in this specimen is marked by muscovite and the feldspar is sericitized.

Specimen 17-98 yielded few, but large, idiomorphic zircon grains with oscillatory zoning (Fig. 14a). Nine of the 22 SHRIMP analyses were removed because they were more than 10% discordant after ²⁰⁴Pb correction. The seven youngest cores analysed in the remaining data yielded a ²⁰⁴Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 398 ± 2 Ma (MSWD = 0.13: Fig. 15a).

Monazite grains from specimen 17-99 were 50–100 µm across, xenomorphic in shape and overgrown by apatite coronas (Fig. 14b, c). Three large monazite grains surrounded by apatite moats and four small xenotime grains (Fig. 14d), and located within the C–S–C′ fabric were selected for in situ geochronology to determine the timing of ductile deformation. Given the texture of the apatite surrounding the monazite and that it forms sigma-shaped clasts within the most altered portions of specimen 17-99, apatite is interpreted to form part of the hydrothermal assemblage of this specimen and as such would date the timing of alteration. Two large apatite grains that form sigma clasts in the most altered portion of specimen 17-99 (Fig. 14e) were selected for in situ geochronology. On electron microprobe maps in wave dispersive mode, monazite is not zoned except for a single core that records zoning in U, Th and Y (Fig. 14c). Neither xenotime nor apatite showed any significant elemental zoning. Muscovite grains from specimen 17-99A defining the C–S–C′ fabrics were picked for ⁴⁰Ar/³⁹Ar geochronology to determine the timing of the brittle–ductile deformation fabrics.

The three large monazite grains were analysed with 61 spots, with the bulk of the data plotting near the concordia curve in a Tera–Wasserburg diagram (Fig. 15b). The analyses defined an older population with a ²⁰⁷Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 410 ± 3 Ma (MSWD = 3.1, n = 9) with several omitted analyses older than 440 Ma, and a younger population with a ²⁰⁷Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 368 ± 10 Ma (MSWD = 0.6, n = 4: Fig. 15b). In addition, the two youngest analyses and the three most discordant analyses also defined a lower intercept age of 352 ± 12 Ma (MSWD = 0.15: Fig. 15b). Four spot analyses were acquired from the two xenotime grains in Figure 14d. These analyses are concordant on a Tera–Wasserburg diagram and yielded a ²⁰⁷Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 365 ± 4 Ma (MSWD = 2.4: Fig. 15c), overlapping the young monazite population. From the 12 apatite analyses in this specimen, 11 were acquired from the grain shown in Figure 14e. On a Tera–Wasserberg diagram, the analyses defined a poorly defined lower intercept age of 376 ± 43 Ma (MSWD = 0.4: Fig. 15d). A one step-heating experiment was carried out on muscovite that yielded a plateau age of 380 ± 2 Ma with 69% of the radiogenic argon released during the heating steps 8–12 (MSWD = 1.77: Fig. 8j). That date is within uncertainty of the younger monazite age and the xenotime and apatite ages for the specimen.

Specimen 17-22 was collected from a mylonitic portion (Figs 11e & 12e) of the 430 ± 2 Ma Taylors Barren granite pluton (U–Pb on zircon: Horne et al. 2003) located to the west of the EHSZ in the CBSZ (Fig. 2b). Twelve monazite grains co-located with muscovite along the foliation (Fig. 14f) were selected for in situ geochronology. The monazite grains were mostly 20–100 µm in diameter, chemically homogenous, xenomorphic and often rimmed by apatite. Hypidiomorphic grains were chemically zoned with a Th-poor core and a Th-rich rim (Fig. 14f). A total of 54 analyses across the 12 grains plot on a Tera–Wasserburg diagram with a cluster of concordant analyses at c. 400 Ma (Fig. 9d). Forty-four analyses comprise a population with a ²⁰⁷Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 387 ± 2 Ma with a high MSWD of 15. Four young analyses defined a ²⁰⁷Pb-corrected ²⁰⁶Pb/²³⁸U weighted mean age of 290 ± 6 Ma (MSWD = 2.5). The younger age group solely comprises analyses from xenomorphic monazite grains, while the older age group is a mixture of spot analyses in both types of monazite grains.

Specimen 17-199 was collected from a pelitic schist (Fig. 11f) in the southeastern flank of the CBSZ (Fig. 2b). In outcrop, the foliation dips steeply to the WNW and the mineral lineation plunges shallowly to the SSW. In thin section, asymmetrical kink bands and C–C′ shear bands indicate a sinistral shear sense (Fig. 12f). Three muscovite aliquots were picked for ⁴⁰Ar/³⁹Ar muscovite geochronology. The step-heating experiments yielded irregular spectra shapes and an integrated age ranging between 378 ± 2 (Fig. 8k) and 377 ± 2 Ma (Fig. 8l).

Our field and thin-section observations are consistent with previous interpretations for the kinematic evolution of the EHSZ (e.g. Lin 1993). While the first deformation event is characterized by SE-side-up kinematics (Fig. 4c), most of the specimens recorded D₂ oblique SE-side-down dextral motion (Figs 4e, f, 11a & 12a, b). We report a widespread altered granite or rhyolite unit bordering the Bras-d'Or flank of the shear zone (Fig. 1b), which is overprinted by C–S–C′ ductile fabrics (Fig. 11b), brecciated (Fig. 11c) and displaying kinematics compatible with D₂ within the shear zone. We further documented a granite mylonite forming the high-strain portion of the CBSZ (Fig. 1b). Thin-section observation of C–C′ fabrics with asymmetrical kink bands (Fig. 12f) indicates oblique sinistral SE-side-up kinematics at odds with the previous kinematic interpretation for the CBSZ (Horne 1995). To determine the timing of deformation of each structure, we dated a variety of units and fabrics with zircon, monazite, xenotime and apatite U–Pb geochronology and with amphibole, muscovite and biotite ⁴⁰Ar/³⁹Ar geochronology (in situ and step heating). In the following subsections we present our interpretations with respect to the deformation events of the EHSZ (D₁, D₂ and breccia) and the CBSZ.

The quartz-diorite orthogneiss (locality 16-30) is located in the Cheticamp Lake orthogneiss of the Aspy terrane, which forms part of the northwestern boundary of the EHSZ. It is cross-cut by a composite dyke of quartz-diorite and pegmatite orientated parallel to the foliation. Although the quartz-diorite is foliated, its pegmatitic core is relatively unstrained, indicating that it either crystallized syn-to post-deformation or that it did not accommodate deformation. Both intrusive phases display chessboard-type quartz subgrains (Fig. 4a, b), which could represent subsolidus or near-solidus strain (Lister and Dornsiepen 1982; Mainprice et al. 1986). The U content and LREE compositions of the zircon from those specimens (Fig. 6) are progressively enriched with respect to the HREEs, which is consistent with progressive crystallization from a common source. Based on these interpretations, dating of these phases would provide constraints on the age of regional deformation, outside the high-strain areas like the EHSZ. The crystallization age of the dyke is best defined by the U–Pb zircon age of 382 ± 2 Ma. The pegmatite specimen, which is likely to represent the younger intrusive phase, yielded an older weighted mean age than the host gneiss, indicating an inherited age, which is frequently observed with pegmatites. The cluster of four zircon analyses with a weighted mean age of 357 ± 7 Ma with an under-dispersed MSWD of 0.16 is likely to represent the crystallization age for this pegmatite and the best estimate on the minimum age of regional deformation (Fig. 16).

The rhyolite (specimen 16-27) is located in the portion of the EHSZ that preserves D₁ SE-side-up kinematic indicators. Because it is the youngest igneous rock that preserves D₁ kinematics, its crystallization age of 424 ± 1 Ma provides a maximum age for D₁ deformation. The ⁴⁰Ar/³⁹Ar amphibole ages of 404 ± 1, 403 ± 3 and 402 ± 3 Ma (specimens 17-38 and 17-107: Fig. 8a–c), which are located in the EHSZ, are interpreted to define the timing of cooling related to the uplift and the exhumation of the shear zone during D₁ (Fig. 16). These results also indicate that the EHSZ may have cooled beyond the amphibole closure temperature (480–578°C: Harrison 1982) during D₁ and that the argon systematics in amphibole remained undisturbed during D₂, at least locally.

The D₂ reactivation of the EHSZ is characterized by oblique SE-side-down dextral kinematics (Lin 1995), which overprinted the Black Brook Granitic Suite to form the triclinic RLSZ (Fig. 2b) (Lin 2001). The D₂ reactivation of the EHSZ was in turn characterized by composite monoclinic kinematic flow (general and transpressive: Piette-Lauzière et al. 2020). The maximum age of D₂ deformation can be interpreted from the 370 ± 1 Ma crystallization age of a granitic dyke (specimen 17-58) that contains D₂ fabrics. The dyke also produced a well-defined in situ⁴⁰Ar/³⁹Ar age (cooling or deformation) of 374 ± 1 Ma, which reinforces the inferred timing of D₂ (Fig. 10a). This interpretation is further strengthened by the monazite and xenotime geochronology results from a muscovite-rich pelitic schist (specimen 16-44A), which also contains D₂ fabrics. Both phases are contained within the muscovite-defined foliation and some xenotime grains are sigma shaped with overgrowth parallel to the foliation (Fig. 7e, f), indicating that they grew, in part, during the D₂ deformation of the specimen. Monazite and xenotime yielded U/Pb crystallization ages of 376 ± 2 and 377 ± 3 Ma, respectively (Fig. 13a, b), undistinguishable from a biotite step-heating age of 375 ± 2 Ma (Fig. 8f). The same specimen yielded slightly younger muscovite dates of 371 ± 2 Ma obtained via in situ and step-heating experiments (Figs 10b & 8d, respectively). Similarly, other specimens (17-76 and 17-56) that present D₂ kinematics yielded a synkinematic xenotime crystallization age of 377 ± 2 Ma (Fig. 9c) and an in situ muscovite ⁴⁰Ar/³⁹Ar deformation or cooling age of 371 ± 2 Ma (Fig. 10c). The muscovite plateau age of 372 ± 2 Ma in specimen 17-56 further reinforces these results (Fig. 8e). Finally, biotite and muscovite grains from specimen 17-51, a little deformed specimen that flanks the shear zone, yielded plateau ages of 370 ± 3 and 376 ± 3 Ma, respectively (Fig. 8h, i). Because all of these mineral systems yield overlapping ages despite different expected closure temperatures (see the Discussion), different fabric intensities and different mineral assemblages, the rocks of the EHSZ are interpreted to have recorded fast cooling during D₂ deformation between 385 and 367 Ma (Fig. 16).

The SE boundary of the EHSZ preserves evidence of D₂ brittle–ductile deformation, associated brecciation and hydrothermal alteration (Fig. 11b–d) (Piette-Lauzière et al. 2018). The U–Pb zircon crystallization age of 398 ± 2 Ma (specimen 17-98) and the older monazite age of 410 ± 3 Ma (specimen 17-99B) are interpreted to reflect the crystallization age of the granite protolith (Fig. 15a, b). Considering the systematic uncertainty related to the methods used (1% for SHRIMP and 2% for LA-ICP-MS), these dates are identical within error to the 403 ± 3 Ma crystallization age of the Cameron Brook Granodiorite (Barr and Raeside 1998), which indicates that these units are likely to be part of the intrusion. In specimen 17-99B, the in situ synkinematic U–Pb monazite age of 368 ± 10 Ma with a xenotime age of 365 ± 4 Ma, the apatite analyses from the breccia matrix with a lower intercept age of 367 ± 43 Ma (Fig. 15b–d), a muscovite ⁴⁰Ar/³⁹Ar plateau age of 373 ± 4 Ma (Fig. 8h) and overall kinematics that are compatible with D₂ indicate that this deformation and potential sericitization was synchronous with late D₂ deformation (Fig. 16). The monazite analyses from specimens 17-58 and 17-99B define similar lower intercept ages of 355 ± 6 and 352 ± 12 Ma, respectively, suggesting that a late fluid flow event may have altered the U–Pb systematics in monazite or triggered the growth of new grains. These lower intercept ages probably best reflect the timing of the brittle deformation of the EHSZ.

In the CBSZ, the monazite grains (specimen 17-22) within the granitic mylonite foliation yielded a U–Pb age of 387 ± 2 Ma (Fig. 9d). Considering the ⁴⁰Ar/³⁹Ar muscovite integrated ages of 378 ± 2 and 377 ± 2 Ma from the phyllite (specimen 17-199: Fig. 8k, l) that is part of the metamorphic suite flanking the CBSZ (Fig. 2b), the monazite age may represent the timing of peak or early retrograde metamorphism preceding the deformation of the CBSZ reflected in the muscovite age. Our field mapping results indicate that the CBSZ was intruded and stitched by the c. 366 Ma Margaree pluton (Fig. 2b) (Sombini dos Santos et al. 2020). These new results demonstrate that the deformation along the CBSZ was at least partially coeval with the reactivation of the EHSZ (Fig. 16).

Our results indicate that the EHSZ formed between c. 420 and 399 Ma during the Acadian Orogeny (D₁: Fig. 16) and was subsequently reactivated in a ductile deformation regime from 385 to 367 Ma (D₂). Fluid flow and brittle–ductile deformation may have continued between 363 and 339 Ma, as indicated by the monazite lower intercept ages from specimens within and neighbouring the shear zone.

This study reported multiple chronometers characterized by different expected closure temperatures yielding overlapping ages: U–Pb in monazite (450–700°C with fluids and 850–1200°C without fluids: Parrish and Tirrul 1989; Gardés et al. 2006; Williams et al. 2011), xenotime (800–1000°C without fluids: Cherniak 2006), apatite (300–600°C: Cherniak et al. 1991), as well as ⁴⁰Ar/³⁹Ar in muscovite and biotite (420–450°C in muscovite and 280–300°C in biotite: Harrison 1982; Cosca and O'Nions 1994; McDougall and Harrison 1999; Harrison et al. 2009). Regardless of the parameters controlling the closure temperature of analysed minerals (e.g. cooling rate, grain size, presence or absence of fluids and lattice damage for argon), the coeval ages reported in this study indicate synchronous deformation and rapid cooling of the rocks of the EHSZ.

The reactivation of the EHSZ was coeval with deformation within the CBSZ at c. 395–369 Ma (Fig. 16). The combined ductile reactivation of the EHSZ with oblique SE-side-down and dextral kinematics with the oblique SE-side-up sinistral ductile deformation along the CBSZ led to the uplift of Aspy in a transtensive environment from c. 385 to 367 Ma (Fig. 17). These results are similar to the exhumation of the Jumping Brook Metamorphic Suite nearby the Western Highlands Shear Zone (Shute 2017) determined by McCarron (2020) based on an apatite U–Pb lower intercept age of 379 ± 3 Ma, and is identical to the ⁴⁰Ar/³⁹Ar deformation or cooling ages determined from hornblende (c. 380–370 Ma), muscovite (c. 375–365 Ma) and biotite (c. 380–368 Ma) in the northern Aspy terrane (Reynolds et al. 1989; Keppie et al. 1992; Price et al. 1999), which also indicates rapid Late Devonian exhumation. This deformation also accommodated the emplacement of the volcano-sedimentary Fisset Brook Formation (373 ± 3 Ma, U–Pb on zircon from a rhyolite: Barr et al. 1995) located along the footwall of the structures flanking the western side of the Aspy terrane. Furthermore, the CBSZ eventually channelled the northern segment of the Margaree pluton at c. 363 ± 2 Ma, as proposed in Sombini dos Santos et al. (2020).

The contemporaneous deformation of the EHSZ and CBSZ, and exhumation of the Aspy terrane, is coeval with the intrusion of the 375 + 5/ − 4 Ma Black Brook Granitic Suite (Fig. 2b: U–Pb on monazite) (Dunning et al. 1990a) and of the 370 ± 4 Ma granite dykes that are parallel to the mylonite fabric in the EHSZ (error adjusted to 1%: specimen 17-58). Thermal weakening of the shear zone combined with the buoyancy of the batholith may have promoted a rapid exhumation of the Aspy terrane and a localization of strain within the pre-existing fabric of the EHSZ (e.g. Cao and Neubauer 2016).

Based on compiled regional deformation ages, the reactivation of the EHSZ and formation of the CBSZ occurred during the Neoacadian Orogeny (Fig. 16). This deformational event includes, at the regional scale, the development of the Cobequid–Chedabucto transpressive shear zone between c. 380 and 337 Ma (Waldron et al. 1989; Pe-Piper et al. 2004, 2017; Bickerton et al. 2018), and the reactivation and change in kinematics of the HDBSZ at c. 389 Ma (D'Lemos et al. 1997; Kellett et al. 2016).

The HDBSZ is characterized by a similar orientation and the Late Devonian dextral strike-slip kinematics of the CCSZ. Moreover, the Ackley Granite that intruded the structure at c. 377 ± 4 Ma (Kellett et al. 2014) yielded overlapping hornblende–biotite ⁴⁰Ar/³⁹Ar total fusion ages of 375 ± 4 and 372 ± 5 Ma (Kontak et al. 1988), respectively, indicating that it may have been rapidly exhumed in a transpressive environment similar to the CCSZ (Archibald et al. 2018). Comparable biotite ⁴⁰Ar/³⁹Ar step-heating ages of 376 ± 4 and 382 ± 2 Ma from intrusions located nearby the HDBSZ indicate that exhumation may have been regional and may have evolved into a brittle deformation regime at c. 351 ± 7 Ma, as determined by K–Ar illite geochronology on the fault gouge (Kellett et al. 2014, 2016).

The results presented herein help to resolve the regional dynamics of the Neoacadian Orogeny during which dominantly transpressive structures were formed in New Brunswick and Newfoundland, synchronously with subordinate transtensive structures such as the EHSZ and the CBSZ. As originally proposed by Murphy and Keppie (1998), this composite architecture helped to accommodate the formation of regional transpressive structures such as the CCSZ but also controlled the opening of the Maritimes Basin. Overall, the kinematics of shear-zone reactivation can be attributed to the oblique convergence of the Meguma terrane and are inherited from the original orientation of the reactivated structures, perhaps similar in setting to the modern-day Dead Sea Transform (e.g. Butler et al. 1998). In part, the obliquity of the collision and the interaction with a Laurentian promontory in Cape Breton Island (Fig. 1) may also explain this change in kinematics along strike.

Here we document the timing of deformation and reactivation of the EHSZ and CBSZ on Cape Breton Island using in situ U–Pb geochronology (zircon, monazite, xenotime and apatite) and a combination of ⁴⁰Ar/³⁹Ar in situ and step-heating experiments (amphibole, muscovite and biotite). The EHSZ records two ductile deformation events: D₁, occurring between 420 and 399 Ma, and D₂, between 385 and 367 Ma. This deformation may have extended into the brittle regime up to c. 339 Ma with brecciation and fluid flow within the EHSZ. The timing of deformation along the CBSZ, located to the west of the EHSZ, is coeval with D₂ deformation at c. 395–369 Ma. The combination of a variety of geochronometers yielding the same age and the overall evolution from ductile to brittle deformation regime during D₂ indicate that the EHSZ accommodated rapid exhumation during its reactivation. Overall, D₂ deformation corresponds to a renewed episode of inboard deformation synchronous with the docking of the Meguma terrane to the composite Laurentian margin characterized by the contemporaneous deformation of the CBSZ and the EHSZ that induced the oblique rapid exhumation of the northern Cape Breton Highlands. This result also indicates that characterizing the system formed by coeval shear zones helps to fingerprint the strain partitioning between transpressive and transtensive structures in accretionary orogens.

We thank Rob Strachan for the handling of the manuscript and two anonymous reviewers and Sébastien Castonguay (NRCan) for their constructive comments. Fieldwork in Cape Breton was supported by James Bridgland and Gena Briand (CBH National Park), and Jean-Philippe and Mary from Wreck Cove. Sandra Barr (Acadia University), Chris White (Nova Scotia Department of Natural Resources), Deanne van Rooyen (Cape Breton University), Gabriel Sombini dos Santos (University of Waterloo), Rob Raeside (Acadia University) and Donnelly Archibald and Caleb Grant (St Francis Xavier University) provided field support. Alfredo Camacho (University of Manitoba), Bill Davis, Nicole Rayner and Greg Case (Geological Survey of Canada), John Cottle, Andrew Kylander-Clark, Francisco Apen, Amy Moser and Alex Johnson (University of California, Santa Barbara) and Mark Button and Sudip Shresta (Fipke Laboratory for Trace Element Research (FiLTER)) helped with geochronology. This is Natural Resources Canada (NRCan) contribution No. 20220396.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

NP-L: conceptualization (lead), formal analysis (lead), investigation (lead), methodology (lead), visualization (lead), writing – original draft (lead); writing – review & editing (lead); KPL: conceptualization (supporting), formal analysis (supporting), funding acquisition (supporting), investigation (supporting), methodology (supporting), writing – review & editing (supporting); DAK: conceptualization (supporting), funding acquisition (lead), investigation (supporting), methodology (supporting), writing – review & editing (supporting); NR: funding acquisition (supporting), writing – review & editing (supporting); JP: investigation (supporting), writing – review & editing (supporting).

Funding for this project was provided by the NRCan Targeted Geoscience Initiative project (TGI-5 and 6), awarded to DAK. Additional funding was provided by a NSERC PGS Doctoral scholarship awarded to NP-L.

The datasets generated during and/or analysed during the current study are available in the Open Science Framework repository: https://osf.io/KEDJP/ [last accessed 16 July 2023].

Specimens 16-30A, 16-30B and 16-30C were crushed, and zircon grains were separated using standard heavy liquids, magnetic separation and hand-picking procedures at the University of California, Santa Barbara. The grains were then mounted in epoxy and polished to expose grain cores. Zircon grains were ablated with a 15–25 µm laser spot from a Cetac 193 nm excimer laser for a duration of 30 s. The ablated material was analysed with a Nu Instruments Plasma HR-ES multicollector inductively coupled mass spectrometer (MC-ICP-MS) for U–Pb isotopes, while trace elements were simultaneously analysed with an Agilent 7700x quadrupole inductively coupled plasma mass spectrometer (Q-ICP-MS) following the laser ablation split stream (LASS) method of Kylander-Clark et al. (2013) with the modification detailed in McKinney et al. (2015). The ablation run included frequent analysis of the zircon reference material 91500 (Wiedenbeck et al. 1995) as the primary calibration standard and zircon GJ1 as a secondary reference material (Jackson et al. 2004) for geochronology. Trace elements were calibrated using NIST 610 glass as a primary reference material. An additional 1.4% uncertainty was added to the ²³⁸U/²⁰⁶Pb ratios to account for calculated dispersion within the analyses of secondary reference materials. No dispersion was added to the ²⁰⁷Pb/²⁰⁶Pb ratios for these specimens.

Additional zircon grains from specimen 16-30B were handpicked in the Fipke Laboratory for Trace Element Research (FiLTER) at the University of British Columbia Okanagan (UBCO) for analysis. The grains were mounted in epoxy and polished to expose grain cores. Zircon grains were ablated with a 25 µm laser spot from a Photon Machines Analyte 193 Excimer laser for a duration of 30 s. The ablated material was then analysed with an Agilent 8900 triple quadrupole ICP-MS. The ablation sequence included several primary (91500: Wiedenbeck et al. 2004) and secondary reference materials (Plešovice: Sláma et al. 2008). An additional 0.7% uncertainty was added to the ²³⁸U/²⁰⁶Pb and ²⁰⁷Pb/²⁰⁶Pb ratios of the unknowns to account for overdispersion measured in the secondary reference material.

Specimens 16-27 and 17-58 were disaggregated at Overburden Drilling Management Ltd (Ottawa) with a Spark 2 Electric Pulse Disaggregator. Heavy minerals were concentrated using a shaking table, and zircon was further concentrated with micro-panning. Zircon grains were handpicked at UBCO, mounted in an epoxy puck at the Geological Survey of Canada geochronology laboratory in Ottawa. The puck was then polished to expose the grain centres, imaged with BSE and cathodoluminescence to assess potential chemical zoning, and coated with gold. The zircon was analysed for U–Pb isotopes using the sensitive high-resolution ion microprobe (SHRIMP II) with an aperture of Kohler 50, which provided a 20–30 µm diameter analytical spot size at the target surface. The unknowns were analysed alongside the primary reference material Temora (Black et al. 2003) and an internal secondary reference material 9910 Rambler (B. Davis and V. McNicoll unpublished data).

The ICP-MS analyses were reduced with iolite (Paton et al. 2011), while the concordia diagrams and weighted means were calculated using the software package ChrontouR (Larson 2020) developed for the open-source R software environment. Different common Pb corrections were used for the data measured on the SHRIMP v. LA-ICP-MS. ²⁰⁴Pb-corrected ²⁰⁶Pb/²³⁸U ratios were used to calculate weighted mean ages for specimens 16-27 and 17-58, which were measured on the SHRIMP-II. A Stacey and Kramers (1975)-based ²⁰⁷Pb correction on the ²⁰⁶Pb/²³⁸U ratios was used to calculate weighted mean ages for specimens 16-30A, 16-30B and 16-30C analysed via LA-ICP-MS. All U–Pb data were plotted on Tera–Wasserburg plots with the uncorrected ratios measured via LA-ICP-MS and the ²⁰⁴Pb corrected ratios measured via the SHRIMP. All ages are reported with their calculated statistic uncertainties and mean square weighted deviation (MSWD). Figure 16, however, is reported with propagated systematic uncertainties of 2% (LA-ICP-MS: Horstwood et al. 2016) and 1% (SHRIMP-II: B. Davis pers. comm. 2019), which reflect long-term reproducibility of the methods and allow for inter-method age comparisons.

Monazite, xenotime and apatite grains were located in samples 16-44A, 17-22, 17-58, 17-76-2, and 17-99B using elemental maps obtained from a tabletop Bruker M4 Tornado micro-XRF at the Laboratoire de microanalyse, Université Laval (Québec). The grains were individually imaged with backscattered electrons on a Tescan Mira 3 XMU scanning electron microscope (SEM) in the FiLTER facility at UBCO. In addition, cathodoluminescence images and element maps were acquired with a Cameca SXFive Field Emission microprobe also located in the FiLTER facility to assess potential chemical zoning. The grains were ablated in situ with a 15–25 µm (monazite and xenotime) and 40 µm (apatite) laser spot at UCSB using the same analytical set-up as employed for zircon analyses there. For monazite and xenotime, the ablation run included frequent analysis of the monazite reference material 44069 (Aleinikoff et al. 2006) as a primary calibration standard and the reference material ‘Stern’ as a secondary reference material for geochronology (Palin et al. 2013) and as a primary reference material for trace element measurements. For apatite, the reference material MAD-UCSB (ID-TIMS age of 467.4 ± 9.8 Ma, 2SE: Apen et al. 2022) was used as a primary calibration reference material. Mount McClure apatite was used as the secondary reference material (Schoene and Bowring 2006). An additional uncertainty of 1.1% was added to the ²⁰⁷Pb/²⁰⁶Pb monazite and xenotime analyses, while an additional uncertainty of 0.25% was added to the apatite ²³⁸U/²⁰⁶Pb ratios to reflect the overdispersion measured in the secondary reference materials.

Duplicate polished thick sections cut from specimens 17-58, 17-56 and 17-44A were imaged with backscattered electrons on a SEM in environmental mode at the Department of Materials Engineering of the University of British Columbia. Muscovite-defined shear bands and C′ shear bands were then cored out at the Geological Survey of Canada (Ottawa) and separated from the glass plate with a solvent. In addition, portions of specimens 17-38B, 17-56, 17-51, 16-44, 17-99A and 17-199 were gently crushed with a mortar and pestle. Amphibole grains for specimen 17-38B, biotite grains for specimen 17-51, and muscovite grains for specimens 17-51, 17-56, 16-44, 17-99A and 17-199 were hand-picked under a binocular microscope. Grains with a diameter of less than 50 µm, with inclusions, discoloration or visible damage were avoided.

The thick section chips and mineral separates were packed into aluminium foil and placed radially inside an aluminium canister along with grains of Fish Canyon Tuff Sanidine (apparent age = 28.201 ± 0.023 Ma, 1σ: Kuiper et al. 2008) so that the lateral neutron flux gradients could be evaluated. The specimens and reference material were irradiated for 78 h in medium flux position 8c using cadmium shielding in the research reactor of McMaster University in Hamilton, Ontario, Canada (160 MWh irradiation).

The irradiated muscovite, biotite and amphibole concentrates from specimens 17-56, 17-51, 16-44, 17-99A, 17-107 and 17-199 were sent to the Geological Survey of Canada for step-heating experiments, whereas the amphibole grains from specimen 17-38B and the thick section chips specimens 17-58, 17-56 and 17-44A were sent to the University of Manitoba for step-heating and in situ experiments.

Upon return from the reactor, single-grain aliquots of muscovite, biotite and amphibole were loaded into individual 1.5 mm-diameter pits in a copper planchet. The planchet was placed in an all-metal, ultra-high vacuum extraction line at the Noble Gas Laboratory of the Geological Survey of Canada, Ottawa. Heating of individual sample aliquots in steps of increasing temperature was achieved using a Photon Machines Ltd Fusion 10.6 55W CO₂ laser. The released Ar gas was cleaned by two hot SAESTM NP-10 getters of St 707 alloy (Zr–V–Fe) held at c. 400°C to remove nitrogen, oxygen, hydrocarbons, water and other active gases, and a room-temperature getter containing HY-STOR^® 201 calcium–nickel alloy pellets to remove hydrogen. The gas was then analysed using a Nu Instruments Noblesse noble gas mass spectrometer, equipped with a Faraday detector and three ion counters. All analyses were run in ion counter multicollection mode (MC-Y mode of Kellett and Joyce 2014). Blank measurements were made after every four unknowns during the analytical session. Mass fractionation and detector efficiencies were determined from repeated measurements of air aliquots throughout the analytical session, whereby ⁴⁰Ar and ³⁶Ar signals were measured on all collectors. ⁴⁰Ar/³⁶Ar ratios were then determined for each collector individually, and for each combination of collectors (excluding ⁴⁰Ar on the Faraday/³⁶Ar on each ion counter). Data reduction and age calculations were performed using Mass Spec software version 7.93 and constants outlined in tables J32–J40. Details regarding data reduction, error propagation and age calculation are summarized in Deino (2001). Corrected argon isotopic data are listed in the Supplementary material. The reported plateau ages satisfy the following criteria: they were derived from three or more consecutive heating steps that were statistically equivalent at the 95% confidence level and comprise more than 50% of the total ³⁹Ar released. Heating steps that released less than 0.5% ³⁹Ar were omitted. MSWD is defined as the mean square of weighted deviates. Neutron flux gradients throughout the sample canister were evaluated by analysing the FCT-SAN sanidine flux monitors included with each sample packet, and interpolating a linear fit against calculated J-factor and sample position. The error on individual J-factor values was conservatively estimated at ±0.6% (2 SE). Because the error associated with the J-factor is systematic and unrelated to individual analyses, correction for this uncertainty was not applied until the ages from isotopic correlation diagrams were calculated (Roddick 1988). Errors in the plateau ages do not include the errors of decay constants. These data were also corrected for ³⁷Ar. Nucleogenic interference corrections were (⁴⁰Ar/³⁹Ar)K = 0.003 ± 0.005, (³⁸Ar/³⁹Ar)K = 0.011 ± 0.010, (⁴⁰Ar/³⁷Ar)Ca = 0.002 ± 0.002, (³⁹Ar/³⁷Ar)Ca = 0.00068 ± 0.00005, (³⁸Ar/³⁷Ar)Ca = 0.00003 ± 0.00003 and (³⁶Ar/³⁷Ar)Ca = 0.00028 ± 0.00016.The decay constant used was λ⁴⁰K_total = 5.463 ± 0.214 × 10⁻¹⁰ a⁻¹ (2 SE) from Min et al. (2000). All ages presented were calculated using an assumed ⁴⁰Ar/³⁶Ar ratio of 298.56 for atmospheric Ar (Lee et al. 2006). All errors are quoted at the 2 SE level of uncertainty. Only the best step-heating experiments are reported in the text, while all the experiments are included in the Supplementary material.

Two single-grain amphibole aliquots and individual grains of the FCT-SAN sanidine flux monitor were placed in a Cu sample tray with a KBr cover slip in a stainless-steel Thermo Fisher Scientific high-vacuum extraction/purification line and baked with an infrared light for 24 h. Individual unknowns were then step heated until fusion using a Photon Machines (55 W) Fusions 10.6 CO₂ laser while the reference materials were fused. After 2 min the reactive gases were removed for both the reference material and the unknowns by two NP-10 SAES getters (one at room temperature and one at 450°C) prior to being admitted to an Argus VI mass spectrometer by expansion. The five argon isotopes were measured simultaneously over a period of 6 min. Measured isotope abundances were corrected for extraction-line blanks, which were determined before every sample analysis. Detector intercalibration (IC) between the different Faraday cups was monitored using Thermo Qtegra software every 2 days by peak hopping ⁴⁰Ar. The intercalibration factor between detector H1 and the compact discrete dynode (CDD) was measured with the unknowns by online analysis of air pipettes (IC values can be found in the Supplementary material). A value of 295.5 was used for the atmospheric ⁴⁰Ar/³⁶Ar ratio (Steiger and Jäger 1977) for the purposes of routine measurements of mass spectrometer discrimination using air aliquots, and in the correction for atmospheric argon in the ⁴⁰Ar/³⁹Ar age calculation. Corrections were made for neutron-induced ⁴⁰Ar from potassium, ³⁹Ar and ³⁶Ar from calcium, and ³⁶Ar from chlorine (Roddick 1983; Renne et al. 1998; Renne and Norman 2001). Data collection and reduction were performed using Pychron software (Ross 2019). The decay constants used were those recommended by Min et al. (2000). Planar regressions were fitted to the standard data, and the ⁴⁰Ar/³⁹Ar neutron fluence parameter, J, was interpolated for the unknowns (see the Supplementary material).

The in situ muscovite geochronology followed a similar method to the amphibole step-heating experiments. While single crystals of reference material were fused using the CO₂ laser, the unknowns were ablated with a 193 nm Excite laser operated at an energy of 7 mJ, for 70 s over an area of 100 × 50 µm. Ablation pits were estimated to be c. 40 µm deep with optical microscopy.