Early Silurian volcanic and subvolcanic rocks are preserved in the lower part of the Chaleurs Group at two locations in northern New Brunswick. At Quinn Point, mafic to intermediate rocks are hosted by sedimentary rocks of the Weir Formation, and at Pointe Rochette, a bed of felsic tuff occurs near the base of the Weir. These rocks are interpreted as the first evidence in New Brunswick of magmatism associated with Late Ordovician – Early Silurian subduction of Tetagouche–Exploits back-arc oceanic crust. At Quinn Point, mafic rocks include a thick basaltic flow or sill and intermediate to mafic cobbles in overlying conglomerate beds. The in situ mafic rocks and the conglomerate clasts are chemically alike and display subduction-related affinities on tectonic discrimination diagrams. At Pointe Rochette, fine-grained felsic tuff contains elevated Th and U and depleted high-field-strength elements, consistent with a subduction-influenced setting, although rare-earth element (REE) abundances are low and the REE profile is relatively flat. A U–Pb (zircon) age of 429.2 ± 0.5 Ma was obtained from the tuff, consistent with the late Llandovery to early Wenlock age of the overlying La Vieille Formation and coinciding with the latter stages of development of the Brunswick subduction complex. Volcanic rocks were emplaced in the arc to arc-trench gap region, probably reflecting local step-back of the magmatic axis due to accretion of continental back-arc ribbons. The low volume of Early Silurian subduction-influenced rocks is probably related to the relatively narrow width of the back-arc basin and the young, “warm” character of back-arc crust.
The northern Appalachian Orogen experienced a protracted and complex tectonic evolution that led to a collage of accreted tectonic elements (van Staal et al. 1998). In simple terms, the bulk of the Appalachian Orogen formed during closure of a proto-Atlantic Ocean (Iapetus), which led to accretion of intraoceanic and continental margin arcs, parts of back-arc basins, and microcontinents. Most of the accreted terranes form a “mobile belt” sandwiched between the Laurentian and Gondwanan cratonic and peri-cratonic domains to the west and east, respectively, and were previously referred to as the Gander and Dunnage zones (Williams 1979; Figs. 1, 2). In northern New Brunswick, the Dunnage Zone is represented by remnants of the Popelogan – Victoria arc and by rocks deposited in the associated Tetagouche–Exploits back-arc basin (e.g., van Staal et al. 2003). In recent years, considerable evidence has been gathered to support the current tectonic interpretation of Dunnage Zone evolution; this interpretation invokes southeast-directed subduction near the Gondwanan margin of Iapetus to explain formation of the arc and back-arc basin in the Middle Ordovician, followed by northwest-directed subduction beneath the Laurentian margin during late Caradocian to late Early Silurian (ca. 425–450 Ma) closure of the back-arc basin (van Staal 1987; van Staal et al. 1990; van Staal 1994; MacLachlan and Dunning 1998; van Staal et al. 1998, 2003, in press).
In northern New Brunswick, evidence of Tetagouche–Exploits back-arc subduction consists of an obducted fragment of Middle Ordovician ophiolite forming part of the Fournier Group (van Staal and Fyffe 1991) and a Late Ordovician – Early Silurian subduction complex (Brunswick subduction complex) characterized by the local development of blueschist (van Staal et al. 1990, in press). Furthermore, a northwest-dipping, “fossil” subduction zone that accommodated middle Paleozoic convergence between Laurentia and Gondwana has been interpreted on reprocessed seismic reflection images in Newfoundland (van der Velden et al. 2004). However, magmatic activity related to this subduction is poorly represented in the rock record of the northern Appalachians at the latitude of New Brunswick, which is probably in part due to the inferred, relatively narrow width of the marginal basin (∼1000 km; van Staal et al., in press). Elsewhere, sporadic Early Silurian continental arc-type plutonic and volcanic activity (e.g., Quimby volcanics) has been documented in the Bronson Hill anticlinorium in southern and northern New England (Tucker and Robinson 1990; Moench and Aleinikoff 2002; Gerbi et al. 2006), in several igneous suites in western Newfoundland (Whalen 1989; Whalen et al. 2006), and in the late Llandovery (~C2–C5) Pointe aux Trembles and Lac Raymond formations in the western Gaspé Peninsula (David and Gariépy 1990; Fig. 1).
Two minor occurrences of mafic–intermediate and felsic igneous rocks have recently been recognized in the lower (Early Silurian) part of the Chaleurs Group in northern New Brunswick, near the unconformable contact with the Fournier Group of the Dunnage Zone. The purpose of this paper is (1) to discuss the chemistry and inferred petrogenesis of mafic and intermediate rocks from the Weir Formation near Jacquet River (Fig. 3); and (2) to describe a thin bed of fine-grained felsic ash tuff that occurs at a slightly lower stratigraphic level at Pointe Rochette (Fig. 3) and document a U–Pb (zircon) isotopic age obtained from the tuff.
Ordovician volcanic and sedimentary rocks of the Popelogan–Victoria arc and Tetagouche–Exploits back-arc basin are unconformably overlain by Late Ordovician to Early Devonian rocks of the Gaspé Belt successor basin (Bourque et al. 1995; Walker and McCutcheon 1995; Wilson et al. 2004; Figs. 1, 2). The Gaspé Belt underlies much of the Gaspé Peninsula, northern New Brunswick, and northern Maine and oversteps the margins of the Humber and Dunnage zones (Malo and Bourque 1993; van Staal and de Roo 1995; Fig. 1). Evidence that the Gaspé Belt is largely underlain by rocks of Dunnage affinity is provided by numerous inliers (Fig. 1) of pre-Late Ordovician volcanic and sedimentary rocks (e.g., van Staal 1994; Winchester and van Staal 1994; Wilson 2003; Schoonmaker and Kidd 2006). In New Brunswick, the Gaspé Belt is inferred to be largely underlain by remnants of the Middle Ordovician Popelogan–Victoria arc and back-arc oceanic rocks of the Fournier Group. Evidence from Ar40–Ar39 dating of low-temperature phengites in the Fournier Group suggests that it was incorporated into the Brunswick subduction complex ca. 440 Ma, i.e., near the Ordovician–Silurian boundary (van Staal et al. 1990, 2003, in press; van Staal 1994; Fig. 2). Subsequent early Llandovery uplift and onlap by younger cover rocks is interpreted as a response to tectonism and arc-trench migration in the Brunswick subduction complex (van Staal et al. 2003).
The Gaspé Belt is classically divided into three subzones, namely (from northwest to southeast) the Connecticut Valley – Gaspé Synclinorium, Aroostook – Percé Anticlinorium, and Chaleur Bay Synclinorium (Rodgers 1970). The study area is located in the northeastern part of the Chaleur Bay Synclinorium (inset in Fig. 3) and is underlain by Silurian rocks of the Chaleurs Group and Early Devonian rocks of the Dalhousie Group that disconformably overlie or are locally in fault contact with the Chaleurs Group (Fig. 3). The Chaleurs Group unconformably overlies the Middle to Upper Ordovician Fournier Group in the Miramichi Highlands and Elmtree Inlier (Alcock 1935; Helmstaedt 1971; Walker et al. 1993a; Walker and McCutcheon 1995; Fig. 3). From oldest to youngest, the Chaleurs Group in this area consists of the Clemville, Weir, La Vieille, Simpsons Field, LaPlante, and Free Grant formations (Walker and McCutcheon 1995).
The Weir Formation comprises texturally and mineralogically immature, fine- to coarse-grained wacke, sandstone, siltstone, and conglomerate (Noble 1976; Webb 1983; Walker and McCutcheon 1995). Igneous rocks have been recognized only in the coastal section at Quinn Point (Fig. 4), but this may be a result of the poor exposure of this formation in general.
Sedimentary structures in the Weir Formation include large-scale dune cross-bedding, ripple-bedding, channeling, irregular lensing of beds, and fining-upward sequences. Sedimentary structures, bedforms, and detailed lithofacies analysis are consistent with fluvial to nearshore depositional environments (Noble 1976; Webb 1983). Provenance studies indicate a “recycled orogenic” source (typical of collisional tectonic settings and subduction complexes) to the south and southeast (Webb 1983). In most places, the Weir Formation unconformably overlies Ordovician rocks of the Fournier Group, although at Pointe Rochette, it conformably overlies greenish grey, fine-grained siliciclastic rocks and minor limestone tentatively assigned to the Clemville Formation (Figs. 3, 5). The Weir Formation is conformably overlain by either the La Vieille Formation or the laterally equivalent Limestone Point Formation; the upper contact is abrupt at Pointe Rochette and Limestone Point but gradational at Quinn Point (Fig. 3).
A Llandovery age is established for the Weir Formation: at Quinn Point, Howells (1975) placed the upper contact of the Weir at the base of a grey “lime sand facies” containing Llandovery C6 brachiopods, whereas Noble (1976) reported Llandovery C3–4 brachiopods in limestone beds near the top of the unit in the same area. The upper age limit of the Weir Formation is constrained by interpreted Llandovery C5–6 to early or middle Wenlockian ages of conodonts and brachiopods from the conformably overlying La Vieille and Limestone Point formations (Berry and Boucot 1970; Howells 1975; Noble 1976; Lee and Noble 1977; Nowlan 1983). At Pointe Rochette, limestone beds of the Clemville Formation underlying the Weir have repeatedly failed to yield age-specific conodonts. However, an isotopic age has been obtained on a newly discovered bed of felsic tuff just below the base of the first Weir conglomerate bed (see the following section).
The Quinn Point section
At Quinn Point, the exposed thickness of the Weir Formation is ∼300 m, measured from the core of a fault-truncated anticline, 500 m west of Quinn Point, to the contact with the overlying La Vieille Formation to the east and west (Fig. 4a). The base of the Weir is not exposed in this section. Massive, granular basalt is observed near the lowest part of the section, i.e., near the faulted core of the anticline, where it appears to be 70 m thick; however, it may actually comprise two thinner layers, as its central part is obscured by unconformably overlying Carboniferous conglomerates (Bonaventure Formation) that occupy a channel in the Devonian (Acadian) paleosurface (Fig. 4b).
Weir sedimentary rocks comprise thin- to medium-bedded, maroon to green or greenish grey, feldspathic and lithic siltstone, sandstone, and pebble–cobble conglomerate. Thick (3 and 10 m) beds of volcanic cobble conglomerate are located about 100 m stratigraphically above (west of) the in situ basalt (Fig. 4b), i.e., near the middle of the exposed Weir section. These conglomerates were evidently emplaced as mass flows based on channeling or loading into underlying beds and comprise well-rounded cobbles and small boulders of mafic and intermediate volcanic rock. Petrographically, mafic clasts in the conglomerate are very similar to the underlying, in situ basalt.
The lack of pillows and massive, non-amygdaloidal nature of the Quinn Point basalt suggests that it may be intrusive (i.e., a sill) rather than extrusive, despite the somewhat glassy groundmass noted in thin section. Lithological and chemical (see the following section) similarities between the in situ mafic rocks and cobbles in the conglomerate beds support derivation of the latter from the former or at least a close genetic relationship. This implies a predominantly extrusive origin for the basalt, but whether emplaced as a flow or sill, the magmatic activity can be constrained to the Early Silurian. The presence of intermediate rocks in the conglomerate but not in situ indicates that the volume and extent of volcanic rocks were initially greater than the current exposure would suggest.
The Pointe Rochette section
At Pointe Rochette (Fig. 5), the Weir Formation consists of ca. 130 m of reddish maroon to light brown or grayish green polymictic, locally quartz-pebble-rich conglomerate. The Weir conformably overlies thin-bedded, non-calcareous siltstone, fine-grained sandstone, and minor limestone and calcareous sandstone assigned to the Clemville Formation, and conformably underlies dark grey lime mudstone of the La Vieille Formation. Just below the contact with the Weir conglomerate, an irregular but generally thin bed (<30 cm) of pink felsic (dacitic) tuff (PR-1) is hosted by greenish grey sandstone. A thin section of the tuff reveals that, despite some recrystallization, good relict vitroclastic textures are preserved, including bubble-wall shards whose presence precludes significant reworking; i.e., it is a primary deposit of juvenile material. The fine grain size, and lack of pumice or lithic clasts, supports deposition in a distal setting following elutriation in a water column. A sample of this tuff was collected for radioisotopic dating to establish absolute constraints on the timing of magmatic activity at its source (see the following section).
Sampling and analytical procedures
Five samples of mafic to intermediate rock from the Quinn Point section and a sample of felsic tuff (PR-1) from the Pointe Rochette section were selected for whole-rock and trace-element analysis (Figs. 4, 5; Table 1). At Quinn Point, two samples of basalt (Q-2 and Q-3) from the massive flow or sill and three mafic to intermediate boulders (Q-1, 5, 6) from conglomerate beds higher in the section were collected. All samples were analyzed at Activation Laboratories Ltd. in Ancaster, Ontario. Analysis for Ni, Cd, Cu, Pb, and Zn were carried out by multi-acid digestion followed by inductively coupled plasma (ICP) spectroscopy; Cr, Sc, As, and Sb were determined by instrumental neutron activation analysis; major oxides, rare-earth elements (REEs), and all other trace elements were analyzed by lithium metaborate–tetraborate fusion and ICP – mass spectroscopy. In mafic rocks, analytical accuracy based on comparisons with internal standards (Lentz 1995) was generally <5% for the major oxides (except for TiO2 (8%) and K2O (18%)), 0%–12% for REE (16% for Lu), and <12% for trace elements used in discrimination or variation plots (Figs. 6–9; except for Rb (21%), Ba (30%) and Zr (20%)). In felsic rocks, accuracy was calculated at <10% for the major oxides (greater for CaO and MnO, which have very low absolute abundances), 0%–12% for REE (16% for Eu), and <12% for trace elements other than Nb (30%) and Ta (46%).
Quinn Point mafic and intermediate rocks
Mafic rocks at Quinn Point contain considerable secondary carbonate, which is reflected in high loss-on-ignition (LOI) and CaO values in the whole rock analyses (Table 1). Normalization of major element data on an LOI-free basis (Table 1) indicates that the bulk composition varies from basalt to basaltic andesite (∼49%–53% SiO2). On a Zr/TiO2 versus Nb/Y discrimination plot, using normalized values for TiO2, the mafic rocks fall along the boundary of the basalt and andesite fields, whereas sample Q-5 from the lower conglomerate bed plots in the alkali basalt field and intermediate rocks from the upper conglomerate bed plot just inside the trachyandesite field (Fig. 6). The three samples from the conglomerate beds are, therefore, more alkalic in terms of their Nb/Y; however, this “alkalic” character is not supported by correspondingly high Zr, TiO2, or P2O5.
Intermediate rocks occur only in the conglomerate beds overlying the massive basalt; the source of these clasts is unexposed in the study area and is possibly completely eroded. However, the presence of these clasts indicates extensive intermediate magmatism in a proximal location, roughly coeval with emplacement of the massive basalt. All five mafic–intermediate rocks have evolved compositions, with Mg#s ranging from 10–12 in intermediate rocks to a maximum of 34 in sample Q-2 (Table 1). However, these low Mg#s are at odds with high Cr and Ni contents in some samples (Table 1), suggesting that considerable replacement of Mg by Ca may have occurred. The abundance of pseudomorphed olivine phenocrysts implies that primary MgO may have been significantly higher prior to olivine alteration.
Carbonate alteration in Quinn Point mafic and intermediate rocks precludes any meaningful characterization based on major elements. They appear to be more evolved than Early Silurian arc-type volcanic and volcaniclastic rocks of the Pointe aux Trembles and Lac Raymond formations (PTLR) in the western Gaspé Peninsula of Québec, which have Mg#s between 50 and 60 (calculated from Table 1 of David and Gariépy (1990)). Obviously, if significant replacement of Mg by Ca has occurred, as suggested previously, the Mg#s of the Quinn Point and PTLR rocks may have originally been more similar.
Caution should be exercised in the interpretation of high field-strength element (HFSE) and REE data in rocks that have undergone carbonate alteration, and HFSE mobility may appear in the form of major changes in ratios involving these elements (e.g., Murphy and Hynes 1986). In the Quinn Point rocks, not all analyses show high LOI and CaO, typical of carbonated rocks, yet there is no systematic variation in HFSE ratios that would support or suggest HFSE mobility in the most carbonated samples. This is also shown by the clustering of the Quinn Point data points on discrimination diagrams (Fig. 7) and is echoed by close parallelism in REE profiles and extended spidergrams (Fig. 8). Trace-element ratios in the Quinn Point samples are very similar to the subduction-related PTLR rocks in the Gaspé Peninsula (David and Gariépy 1990; Fig. 7). The Quinn Point mafic rocks show substantial enrichment in Th compared with Nb (Fig. 7a) and fall within (or near) the subduction-related basalt fields on Th–Hf–Ta and Zr–Ti–Y diagrams (Figs. 7b, 7c). In contrast to these arc-type signatures, they plot within or near the within-plate-basalt fields on Zr–Nb–Y and Zr/Y versus Zr diagrams (Figs. 7d, 7e). On a Th/Yb versus Ta/Yb diagram, the Quinn Point rocks display enrichments in both Th, typical of arc volcanic rocks, and Ta, characteristic of within-plate basalts, although elevated Th/Ta (12.5–22.0) is consistent with an arc setting (e.g., Pearce 1982; Fig. 7f).
Despite some evidence of within-plate affinity, as mentioned previously, the Quinn Point rocks are chemically distinct from northern New Brunswick basalts that display clear within-plate signatures, namely the late Early Silurian Bryant Point Formation (Dostal et al. 1989), and alkalic and tholeiitic basalts of the Ordovician Bathurst Supergroup (van Staal 1987; van Staal et al. 1991; Wilson et al. 1999; Fig. 7). A clear distinction is also evident on chondrite-normalized REE plots and primitive-mantle-normalized spidergrams; e.g., the REE profiles of the Quinn Point and PTLR rocks have steeper slopes and greater enrichment in light REE (LREE) compared with within-plate basalts (Fig. 8a). However, the Quinn Point rocks do not display the flat heavy REE (HREE) profile, typical of arc-type basalts (e.g., Pearce et al. 1995). The depletion in Yb, Lu, etc. implies crystallization of phases with high bulk partition coefficients for the HREE. On extended trace-element plots, the Quinn Point and Quebec samples display jagged patterns with negative Nb and Ti anomalies and positive spikes in Th and LREE, typical of subduction-influenced basalts (Fig. 8b). Both suites are different from the within-plate pattern exemplified by the Bryant Point Formation and Bathurst Supergroup (Fig. 8b).
Pointe Rochette felsic tuff
Fine-grained vitric ash tuff at Point Rochette has a rhyodacitic composition (Fig. 5; Table 1). Considerable mobility of alkali oxides and large-ion-lithophile elements is indicated by very low K2O, Cs, and Rb abundances and high Na2O. On Rb versus Y + Nb and Nb versus Y discrimination diagrams (Figs. 9a, 9b), the Pointe Rochette tuff plots in the volcanic-arc field. The Pointe Rochette tuff is clearly distinct from within-plate rhyolites of the Late Silurian (Ludlovian–Pridolian) Benjamin Formation (Chaleurs Group) (cf. Dostal et al. 1989). Compared with the Benjamin Formation, the Pointe Rochette tuff has much lower absolute abundances of all REE, a flatter profile (LaN/YbN = 2.8, Table 1), and a much less prominent negative Eu anomaly (Fig. 9c). Low REE abundances, especially in the LREE, appear to be inconsistent with a subduction setting. Furthermore, across-the-board depletion of REE (except for Eu) is difficult to explain by crystallization of REE-bearing phases such as apatite or hornblende, as variations in partition coefficients should be reflected in significantly greater or less depletion in specific REE. It is suggested that low REE abundances can be explained by syn-depositional winnowing of REE-bearing phases in a water column.
All of the foregoing chemical traits are illustrated on a primitive mantle-normalized, extended trace-element spidergram (Fig. 9d). Depletion in Cs, Rb, Ba, and K can be explained by post-depositional alkali mobility and REE depletion by surmised syn-depositional processes. Note that a negative Nb anomaly has been obscured by the anomalously low abundance of K (Fig. 9d). Among the remaining elements, enrichment in Th and U and depletion in HFSEs (Nb, Hf, and Zr) are consistent with a subduction-influenced setting.
Sample PR-1 of fine-grained ash tuff was dated by isotope dilution thermal ionization (ID-TIMS) analysis at the Jack Satterly Geochronology Laboratory, University of Toronto, Ontario. Zircon analyses 1–4 (Fig. 10) were annealed and leached to ensure removal of radiation-damaged zones that may have lost Pb, as described by Mattinson (2005) and tested and modified by Mundil et al. (2004). Grains were placed in a muffle furnace at ∼950 °C for 60 h to anneal-damaged lattice sites, followed by leaching in hydrofluoric acid (HF) in Teflon dissolution vessels at 200 °C for 17 h. Zircon analyses 5–7 were air abraded (Krogh 1982). Zircons were weighed on a microbalance and should be accurate to about ±3%–5%, which affects only U and Pb concentrations and not the age information. The grains were washed prior to dissolution and a 205Pb–235U spike was added to the Teflon dissolution capsules during sample loading. Zircon was dissolved using ∼0.10 mL of concentrated HF and ∼0.02 mL of 7N HNO3 in teflon bombs at 200 °C (Krogh 1973) for 5 days and redissolved in ∼0.15 mL of 3N HCl. U and Pb were isolated from the zircon solutions using 50 μL anion exchange columns and deposited onto outgassed rhenium filaments with silica gel (Gerstenberger and Haase 1997). Pb was analyzed with a VG354 mass spectrometer using either a Daly collector in pulse-counting mode or multi-dynamic mode. All common Pb was assigned to procedural Pb blank. U was measured by using the axial Faraday or axial Daly collector in pulse-counting mode. Dead time of the measuring system for Pb is 22.8 ns and 20.8 ns for U. The mass discrimination correction for the Daly detector is constant at 0.07%/atomic mass unit (AMU). Amplifier gains and Daly characteristics were monitored using the SRM982 Pb standard. Thermal mass discrimination corrections are 0.10%/AMU.
Decay constants are those of Jaffey et al. (1971). All age errors quoted in the text and table and error ellipses in the concordia diagrams are given at the 95% confidence interval. Plotting and concordia age calculations are from Isoplot 3.00 (Ludwig 2003).
U–Pb data for five of seven single zircons are plotted on a concordia diagram (Fig. 10), and the data for all samples are presented in Table 2. Analyses 1–5 are concordant and overlap within their analytical uncertainties; however, data for 4 and 5 plot slightly below analyses 1–3 and may have lost a small amount of Pb. Thus, the most robust age interpretation is the weighted mean 206Pb–238U age of analyses 1–3, which is 429.20 ± 0.47 Ma (mean square weighted deviate = 0.3). All three analyses were obtained from zircons that were annealed and leached prior to analysis. The zircon for analysis 4 was similarly pretreated while those for analyses 5–7 were air abraded. The two points that are not plotted (analyses 6 and 7) are distinctly older zircons inherited from the magma source or entrained during eruption. These grains have 207Pb–206Pb ages of 1118 ± 4 Ma (5% discordant) and 961 ± 6 Ma (5% discordant).
The Pointe Rochette tuff is separated from the overlying La Vieille Formation by a thickness of ca. 130 m of Weir conglomerate. Although no estimate of the time interval represented by this thickness of conglomerate is possible, the coarse grain size and mineralogical and textural immaturity of these beds support rapid deposition. Assuming the Llandovery–Wenlock boundary lies at ca. 428 Ma (Gradstein et al. 2004), the 429.2 Ma age of the Pointe Rochette tuff is consistent with the late Llandovery – early Wenlock age of the La Vieille Formation.
Late Ordovician to Early Silurian sedimentation in the Gaspé Belt occurred in an upper plate (fore-arc) setting with respect to inferred northwest-directed subduction of Tetagouche–Exploits back-arc basin lithosphere. Fore-arc magmatism, represented by the Quinn Point and probably also the Pointe Rochette volcanic rocks, is a rare phenomenon and only occurs in special tectonic circumstances (e.g., Macdonald et al. 1999). Given the composition of the magmas and established tight constraints on the regional tectonic setting and evolution (van Staal 1994; van Staal et al. 2003) the following three mechanisms can be considered: (1) arc-trench migration induced by slab rollback and (or) stepping back of the subduction zone due to accretion of large unsubductable continental back-arc ribbons (e.g., the large Tetagouche block, van Staal et al. 2003), such that the arc’s magmatic axis locally moved into the fore-arc region; (2) a pause in subduction (Macdonald et al. 1999); and (3) slab breakoff (e.g., Whalen et al. 2006). The time of Salinic collision between the Brunswick subduction complex and the Gander margin in New Brunswick and hence time of slab breakoff is best constrained to the Wenlockian (Fig. 11; van Staal et al., in press), mainly because collision probably postdates deposition of the La Vieille Formation and predates or is coeval with deposition of Ludlovian–Pridolian sedimentary rocks of the Gaspé Belt (van Staal and de Roo 1995; Dimitrov et al. 2004; Wilson et al. 2004). A Wenlockian collision is supported by 39Ar/40Ar dating of collision-related S2 phengite in the Brunswick subduction complex (van Staal et al. 2003). The Llandoverian (Telychian) age of the Quinn Point volcanic rocks thus suggests that they could be related either to arc migration or a temporary cessation of subduction. We favour the first option because there is no evidence for the latter process at this time in the Brunswick subduction complex (van Staal et al. 2003). Weak within-plate chemical signatures in these volcanic rocks (e.g., high Zr, Ta, and low V) may reflect their generation near a major, leaky(?) transform fault that linked the St. Lawrence promontory with the Quebec reentrant (Thomas 1977; Stockmal et al. 1990; Malo and Kirkwood 1995). Local transtension associated with this transform could have created a favourable environment for magmas to reach the surface.
The 429 Ma age of the Pointe Rochette ash, very near the Llandovery–Wenlock boundary, allows it to be related to magmatism generated during slab breakoff, but given its subduction-related signature and dissimilarity to known within-plate rhyolites in the region, we tentatively interpret it as the “last gasp” of calc-alkaline continental arc volcanism (Fig. 11). If so, slab breakoff took place shortly after 428 Ma, during the Wenlockian and (or) Ludlovian. This conclusion is significant because slab breakoff in the same tectonic setting along strike in Newfoundland is thought to have taken place 4–6 Ma earlier (Whalen et al. 2006). Our interpretation thus suggests that (1) breakoff was diachronous and becoming younger towards the southwest, and hence (2) Salinic collision-related deformation also becomes younger to the southwest within the Quebec reentrant, which was predicted by Malo and Kirkwood (1995). In spite of this, the rapid progression in northern New Brunswick from arc-type magmatism at 429 Ma to within-plate magmatism ca. 423 Ma (i.e., following deposition of the late Llandovery – early Wenlock La Vieille Formation), closely mirrors that documented in Newfoundland by Whalen et al. (2006).
The scarcity of arc volcanic rocks in the Lower Silurian rock record of the modified Laurentian margin is probably related to the narrow width (∼1000 km) of the Tetagouche–Exploits back-arc basin and to the relatively young (ca. 20 Ma), “warm” character of the back-arc oceanic crust during Silurian subduction (van Staal 1987, 1994; van Staal et al. 1998, in press; Whalen et al. 2006). Arc volcanic rocks are often sparse or even absent above subduction zones characterized by young slabs (Simkin and Siebert 1984; Kirby et al. 1996) because under these conditions metamorphism-driven slab dehydration primarily occurs at relatively shallow depths beneath the fore arc instead of under the asthenospheric mantle wedge. Hence, fluids do not flux the asthenospheric wedge as in cold downgoing slabs and do not facilitate typical arclike melting.
It is significant in this context that Kirby et al. (1996) pointed out that subduction-related magmas in these settings may have unusual compositions, including anomalous REE profiles. In the Quinn Point rocks, Yb (and other HREE) contents are well below average crustal values of 2–3 ppm (e.g., Henderson 1984; Rudnick and Fountain 1995), which is difficult to explain if conventional petrogenetic processes are assumed. Characteristics of the Quinn Point rocks such as HREE depletion, high La/Yb, low Ba/La, high Al2O3 (>15%), and slight positive Eu anomalies are shared by so-called adakitic magmas, whose origin has been related to slab melting (Kay 1978; Defant and Drummond 1990; Peacock et al. 1994; Kay and Kay 2002). One may speculate, therefore, that dehydration metamorphism of the young, warm subducting slab was followed by partial melting of a garnet-bearing eclogitic residue. However, other scenarios have been proposed as the source of adakitic signatures, such as melting at the base of thickened orogenic crust (Kay and Abruzzi 1996; Kay and Kay 2002) or entry of a crustal component into the mantle wedge via subduction erosion (von Huene and Ranero 2003; Kay et al. 2005). Both have potential application in the present case: the former by structural underplating and stacking of thrust nappes in the Brunswick subduction complex (Fig. 12) and the latter by the erosive potential of relatively high-relief continental back-arc ribbons entering the subduction zone.
Mafic to intermediate volcanic–subvolcanic rocks hosted by the Weir Formation are interpreted as the first evidence of Early Silurian subduction-related magmatism in northern New Brunswick. Although the lithogeochemical signatures of these rocks do not allow a unique interpretation by themselves, a subduction interpretation is more consistent with other geological evidence (e.g., low-temperature – high-pressure metamorphic rocks) that constrain the geodynamic setting of the northern Appalachians during the Early Silurian (van Staal et al. 1990, 1998, in press; van Staal 1994). In addition, the trace-element data show a close resemblance to mafic to intermediate, Lower Silurian subduction-related volcanic rocks of the Pointe aux Trembles and Lac Raymond formations in Quebéc (David and Gariépy 1990).
A thin bed of felsic ash tuff at Pointe Rochette also exhibits subduction-influenced chemical signatures, although it may represent volcanic activity at a more distal location. The tuff has yielded a U–Pb (zircon) age of 429.2 ± 0.5 Ma, essentially coeval with the latest deformation and metamorphism in the Brunswick subduction complex as determined by Ar40/Ar39 dating of low-temperature phengites (van Staal et al. 2003). Hence, late Llandovery volcanic rocks in the Chaleurs Group appear to manifest “last-gasp” subduction volcanism prior to slab breakoff and initiation of within-plate volcanism in the Wenlock. Comparison with the timing of slab breakoff in the same tectonic setting in Newfoundland (Whalen et al. 2006) suggests that breakoff, uplift, and Salinic collision-related deformation all occurred 4–6 Ma later to the southwest in the Quebec reentrant, i.e., northern New Brunswick.
The presence of abundant intermediate volcanic cobbles and boulders in conglomerate beds at Quinn Point implies that effusive rocks were more abundant and compositionally diverse than the present exposure would indicate. Nevertheless, Lower Silurian arc volcanic rocks are rare in the northern Appalachians, possibly because of the young, warm character of Tetagouche–Exploits back-arc lithosphere. The fore-arc location of the Quinn Point and Pointe Rochette volcanic rocks is likely a result of arc-trench migration induced by stepping back of the subduction zone following accretion of microcontinental blocks in the back arc (van Staal et al. 2003). Some chemical signatures, e.g., HREE depletion, high La/Yb, high Al2O3, and slight positive Eu anomalies, are similar to those seen in adakites, suggesting that the parent magma may have been derived by partial melting of a metamorphosed (eclogitic) subducting slab or other processes involving partial melting in the presence of residual garnet.
We thank J. Whalen and S. McCutcheon for critical reading of early versions of this manuscript. Peer reviews by P. Pilote and B. Murphy, and recommendations by Associate Editor N. Pinet have significantly improved the manuscript.
Geological Survey of Canada, Contribution No. 2008-0166.
Contribution to Natural Resources Canada’s Targeted Geoscience Initiative 3 (2005–2010).
- Received January 18, 2008.
- Accepted September 17, 2008.
- Published on the NRC Research Press Web site at http://cjes.nrc.ca on October 24, 2008.