Introduction

Numerous early Mesozoic continental tholeiitic basalt flows and dykes are found along the eastern margin of North America from Florida and South Carolina to Nova Scotia and Newfoundland and form part of the Newark Supergroup. The Newark Supergroup is composed of Late Triassic–Early Jurassic (Hettangian) fluvial and/or lacustrine sedimentary rocks (Froelich and Olsen 1985) that are intercalated with basaltic flows. The emplacement of the basaltic rocks occurred at ~200 Ma and lasted <1 million years (Olsen 1997; Olsen et al. 1998; Marzoli et al. 2011). The presence of a long, continuous geophysical reflector off the continental margin of North America indicates a much more extensive offshore continuation of these basalts. The basaltic rocks are part of the larger, Central Atlantic Magmatic Province (CAMP) that is exposed along the continental margins of South America, Western Europe, and West Africa (Fig. 1). The volcanic and plutonic rocks were emplaced along with sedimentary rocks within extensional basins during lithospheric extension and continental rifting that preceded the opening of the Atlantic Ocean (Dostal and Durning 1998; Marzoli et al. 1999; Pe-Piper and Reynolds 2000; Whithjack et al. 2012).

Fig. 1
figure 1

Based on Merle et al. (2013)

Distribution of volcanic and plutonic rocks of the Central Atlantic Magmatic Province and location of the Fundy Basin.

Full size image

The CAMP is one of the largest flood basalt provinces, covering an area ≥107 km2, and was contemporaneous with the end-Triassic mass extinction (Marzoli et al. 1999, 2004; Whiteside et al. 2010; Blackburn et al. 2013). The compositional variation of the mafic volcanic rocks, along with the radial nature of dykes around an inferred volcanic epicenter, is considered to be strong evidence in favor of a mantle plume origin but there remain compelling arguments against such a model (Greenough and Hodych 1990; Wilson 1997; Janney and Castillo 2001; Storey et al. 2001; Cebriá et al. 2003; Melankholina and Sushchevskaya 2015). Specifically, the thermal regime of the mantle that generated the flood basalts was not anomalously hot; there is no hotspot track, and that rifting began ~25 million years before the eruption of the basalts (McHone 2000; Callegaro et al. 2013; Hole 2015; Whalen et al. 2015).

One of the defining characteristics of CAMP basaltic rocks from the eastern North America (ENA) sub-province is the spatial-compositional variation from South to North (Weigand and Ragland 1970; Cummins et al. 1992). The basaltic rocks can be subdivided into three principal groups: (1) olivine-normative, (2) high-TiO2 quartz-normative, and (3) low-TiO2 quartz-normative with a less common high-Fe2O3t quartz-normative type. The olivine-normative rocks predominate within the southern ENA (SENA), whereas the quartz-normative rocks are predominant in the northern ENA (NENA). The transition between the olivine-normative and the quartz-normative North occurs around Virginia–North Carolina (Weigand and Ragland 1970) although some olivine- or quartz-normative rocks are found outside their regions. The spatial-compositional variability is attributed to a number of processes including: mantle source (enriched subcontinental lithospheric mantle or mantle-plume), open- versus closed-system magmatism, depth and/or density-controlled partial melting and fracture-zone or transform fault influence (Weigand and Ragland 1970; Greenough and Hodych 1990; Cummins et al. 1992; Puffer 1992; Murphy et al. 2011; Callegaro et al. 2013, 2014; Merle et al. 2013; Whalen et al. 2015).

The North Mountain basalt (NMB) erupted within the Fundy Basin of the Northern Appalachians and is one of the best exposures of CAMP-related rocks in Eastern North America. The rocks are well studied and dated however unlike other regions of the CAMP the primary melt composition and thermal regime of the NMB has not been investigated. In this paper, we present new whole rock geochemical data from the North Mountain basalt. Samples were collected from surface exposures along the entire length of the formation as well as from a 160-m-deep drill well. We use the data to estimate the primary melt composition of the NMB and their mantle potential temperatures (T P) in order to compare with basalt from other regions of the ENA sub-province and constrain the extent of mantle source heterogeneity between the SENA and NENA.

Geological background

The North Mountain basalts were emplaced in the Bay of Fundy graben, the most northerly of the sixteen basins of the Newark Supergroup (Schlische et al. 2002) that run parallel to the continental margin of the North America (Figs. 2, 3). The first descriptions of the North Mountain basalts were produced by Powers (1916), whereas Powers and Lane (1916) reported the evidence for differentiation in these basalts which were used by Bowen (1916) to document his model of differentiation in mafic magmas. Moreover, some of the basalt flows appear to have experienced silicate liquid immiscibility (Greenough and Dostal 1992a; Shellnutt et al. 2013).

Fig. 2
figure 2

Modified from Dostal and Greenough (1992)

Geological map of the North Mountain basalt and sampling locations. Inset map of Atlantic Canada showing the distribution of CAMP-related mafic volcanic and intrusive rocks. 1 North Mountain Basalt, 2 Shelburne Dyke, 3 Caraquet Dyke, 4 Avalon Dyke.

Full size image
Fig. 3
figure 3

Field photographs of the NMB. a Columnar joint structure of the upper flow (Long Island, near Tiverton). The leftmost column is ~9 m high (photo taken by D. Kontak). b Coastal exposure of the upper flow beneath the Margaretsville lighthouse. c North Mountain basalt overlying the Fundy Group (Triassic–Jurassic) sedimentary rocks of the Blomidon Formation at Five Islands Provincial Park (photo taken by D. Kontak). d Plan view of columnar joints along the shore line at Canada Creek. The pen, near center of photo, is ~15 cm in length. e Middle flow basalt with amygdules (green, white, red, orange), West of Harbourville. f Xenolith of middle flow basalt with amygdules within the upper flow (French Cross, Morden). The pen is the same in d

Full size image

The NMB conformably overlies Early Mesozoic siltstones and shales (Blomidon Formation) and continental red conglomerates and sandstone (Wolfville Formation) of the Fundy Group. The Mesozoic strata lie unconformably on Carboniferous and older rocks. The basalts which were emplaced just above the Triassic–Jurassic boundary are unconformably overlain by lacustrine limestones and continental clastic sedimentary rocks (Olsen et al. 1987; Cirilli et al. 2009). The Mesozoic rocks of the Bay of Fundy form a large, asymmetrical, plunging syncline that dips more-steeply on the northern side. The northern boundary of the basin is delineated by a major south-dipping fault, which may represent the western extent of the boundary between the Meguma and Avalon terranes. Radiometric dating (U–Pb, 40Ar/39Ar) of basalts yielded an age of 202 Ma (Hodych and Dunning 1992; Kontak and Archibald 2003) and Olsen et al. (1982, 1987) assigned them to the earliest Jurassic on the basis of stratigraphy and paleontology.

The basalts form a prominent cuesta which extends for about 200 km along the southern shore of the Bay of Fundy with correlative flows on the north shore of the Minas Basin and in Cape Breton Island (Greenough and Dostal 1992b; White et al. 2017). The basalts thin from southwest (~400 m thick) to the northwest (~275 m), probably underlie most of the Bay of Fundy and cover an area of about 10,000 km2 (Dostal and Greenough 1992). The basalts underwent zeolite facies metamorphism (Aumento 1966), which modified the primary mineralogy of the middle unit. The altered rocks have high contents of LOI (up to >5 wt%). The alteration led to a redistribution of alkalis and Ca in some samples (Dostal and Dupuy 1984). In addition, Cu was also affected during metamorphism. Dostal and Dupuy (1984) documented that while this element is depleted in many samples, some altered basalts have rather high concentrations (>1000 ppm Cu).

Petrography

The basaltic formation includes three units. The lower and upper units are composed of thick medium- to coarse-grained massive flows, whereas the middle unit up to 50-m thick consists of a series of thin extensively altered amygdaloidal lava flows with abundant zeolites and quartz, which also fill abundant amygdules. The basalts of the lower and upper units are fresh and contain zoned microphenocrysts of plagioclase (An60–80) and augite as well as minor Fe–Ti oxides set in a matrix composed of plagioclase, augite, pigeonite, Fe–Ti oxides, accessory apatite, and devitrified glass. The NMB has long been known for its varied and abundant zeolite minerals. In fact, it hosts the type locality for the mordenite and a comprehensive summary of their occurrences was published nearly 100 years ago (Walker and Parsons 1922). However, the degree of zeolitization of the basalts is variable.

Sampling and analytical methods

A total of forty samples were collected for this study. The twenty-four samples (11,392 to 11,420) are specimens collected from outcrops of the basaltic belt (cuesta), in area between Centerville (near Kentville) and Brier Island whereas sixteen samples labeled 13 to 690 are the core samples from the diamond drill hole GVA-77-3 drilled ~1 km southwest from the village of Morden in the Annapolis Valley (Fig. 2). The numbers refer to the depth from the surface (in feet). The drill hole was described by Kontak et al. (2005). From the surface to the depth of about 520 feet, the rocks are from the middle unit which is composed of 15 flows each about 3–20 m thick. The samples numbered 537.5–690 are from the lower unit (Table 1).

Table 1 Geochemical data of the North Mountain basalt
Full size table

Whole-rock abundances of major and some trace elements (Rb, Ba, Sr, Zr, Y, Ga, Cr, Ni, and V) were determined on glass disks and pressed pellets, respectively, using a Philips PW1400 X-ray fluorescence spectrometer at Saint Mary’s University, Halifax (Dostal et al. 1994). The analytical uncertainties were estimated to be generally ~1% for the major elements and 5–15% for the trace elements. Rare earth and the other trace elements (Th, Nb, Hf) were determined by inductively coupled plasma-mass spectrometry (ICP-MS) at the Geoscience Laboratories of the Ontario Geological Survey. Precision and accuracy are given by Ayer and Davis (1997) and are generally within 5–10%. The full data suite is listed in Table 1.

Results

The North Mountain basalts are typically hypersthene-normative continental tholeiites with pigeonite in their mode (Fig. 4a, b). Overall, their composition resembles continental flood basalts from other continental flood basalt provinces and there does not appear to be significant compositional differences among the three units (upper, middle, lower) except for the extent of alteration (Dostal and Dupuy 1984; Dostal and Greenough 1992). The rocks from this study are similar to the basalt from the NENA in general and have trace element ratios indicative of an enriched mantle source that likely had a reducing relative oxidation state (Fig. 4c, d). The basalts display noticeable variations in major element compositions with Mg# (molar Mg/Mg + Fetotal) in our samples ranging from 0.68 to 0.34 and with MgO (wt%) from 9.5 to 3.7 indicating that many rocks experienced extensive fractional crystallization. The samples with the highest Mg# have the lowest incompatible trace element concentrations including Zr but the highest contents of pyroxene phenocrysts and Cr and Ni. The pegmatitic basalts have the lowest Mg# and MgO but high Zr (>150 ppm in Fig. 5).

Fig. 4
figure 4

a Na2O + K2O (wt%) versus SiO2 (wt%) chemical classification of volcanic from the ENA sub-province of the CAMP and the North Mountain basalt. F foidite, P picro-basalt, B basalt, Ba basaltic andesite, A andesite, D dacite, R rhyolite, Td trachydacite, T trachyte, Ta trachyandesite, Bta basaltic trachyandesite, Tb trachybasalt, TB tephorite or basanite, PT phono-tephrite, Tp tephriphonolite, Ph phonolite. b A (Na2O + K2O wt%)–F (FeOt wt%)–M (MgO wt%) diagram showing the tholeiitic trend of the ENA basalt and North Mountain basalt. c Th/Yb versus Ta/Yb basalt discrimination diagram of Wilson (1989) showing the differences between subduction and oceanic basalts derived from depleted and enriched source. Vectors show influence of each component, S subduction component, C crustal component, W within plate enrichment, f fractional crystallization. d Diagram indicating the redox state of the ENA basalt and North Mountain basalt using the bulk-rock V/Ga ratio. Reference lines at various fO2 are after Mallmann and O’Neill (2009). ENA data from Chowns and Williams (1983), Grossman et al. (1991), Pe-Piper and Reynolds (2000)

Full size image
Fig. 5
figure 5

Plots of Zr (ppm) versus a Al2O3 (wt%), b TiO2 (wt%), c P2O5 (wt%), d Cr (ppm), e Th/La, and f Dy/Yb of the ENA basalt and North Mountain basalt. The Th/La plot depicts the vectors of crustal contamination and the Dy/Yb plot depicts the vectors of melting depth. ENA data from Chowns and Williams (1983), Grossman et al. (1991), Pe-Piper and Reynolds (2000). CC crustal contamination, FC fractional crystallization, MD melting depth

Full size image

On the primitive-mantle normalized trace element patterns, the North Mountain basalts exhibit shapes which are common in continental flood basalts. They display negative Ba, Nb, Sr, and Ti anomalies (Fig. 6a). The negative Ba and Sr anomalies are likely caused by fractional crystallization of plagioclase in the magma chamber or during magma ascent. The chondrite-normalized REE patterns (Fig. 6b) are subparallel, have moderately sloping with (La/Yb)n ~3–4, consistent with the low-pressure fractionation dominated by pyroxenes and plagioclase. The patterns do not have noticeable Eu anomalies despite the presence of plagioclase phenocrysts and Sr anomalies in the primitive mantle normalize plot.

Fig. 6
figure 6

a Primitive mantle normalized incompatible element and b chondrite-normalized rare-earth element plots of the North Mountain basalt. Normalizing values of Sun and McDonough (1989)

Full size image

Discussion

Petrogenesis of the North Mountain Basalt

Chemically, the NMB are typical continental tholeiitic basalts with characteristic Fe–Ti-enrichment fractionation trend and are comparable to the high-TiO2 quartz normative type of Weigand and Ragland (1970). An increase of Fe, Ti, P, and V and decrease of Mg, Ca, Cr, and Ni with decreasing of Mg# but increase of Zr (taken in Fig. 4 as an index of fractionation in lieu of Mg# to avoid the alteration effect) is typical of tholeiitic fractionation trends pointing to crystallization of pyroxenes and plagioclase but a negligible fractionation effect of Fe–Ti oxides during the differentiation (Fig. 4b). The steeper decrease of Cr relative to Ni and Co in the basalts with higher Mg# suggests clinopyroxene crystallization but argue against significant olivine fractionation. However, the low contents of Ni in most samples imply a crystallization of olivine in a differentiation stage prior to eruption. Moreover, the slight increase of Al, Sr, and Al/Ca ratios in less fractionated basalts (<150 ppm Zr; Fig. 5) indicates a predominance of clinopyroxene over plagioclase during differentiation.

The variation of La/Yb versus Yb in basaltic rocks have been used to differentiate between compositional changes due to fractional crystallization, differences in degree of melting and source heterogeneity (Fan et al. 2008; Dostal et al. 2016). The La/Yb and Yb trend shown in Fig. 7 is consistent with fractional crystallization (horizontal trend) and argues against large-scale source heterogeneity in the ENA basalts, although some NENA basalts follow the source heterogeneity trend. Several incompatible element ratios are sensitive to melting conditions such as pressure or depth of melting. The Dy/Yb versus Zr plot (Fig. 5f) suggests that all parent melts were generated at about the same depth in the spinel peridotite stability field (60–70 km depth).

Fig. 7
figure 7

ENA data from Chowns and Williams (1983), Grossman et al. (1991), Pe-Piper and Reynolds (2000)

La/Yb versus Yb (ppm) variation diagram for North Mountain basalts and the ENA. Vectors for fractional crystallization (FC), increase of the degree of melting, and source heterogeneity are after He et al. (2010). PM partial melting and heterogeneity = source heterogeneity.

Full size image

The elevated Th/La ratios are commonly used as indices to identify crustal contamination (Jochum et al. 1991; Dostal et al. 2016). Constant Th/La ratios with Zr (Fig. 5e) suggest that there was no significant shallow-seated crustal contamination of the basalts, although relatively high values (0.15–0.20) compared to the mantle values (0.12—Sun and McDonough 1989) indicate that the parent magma was either contaminated prior to emplacement or that it was derived from subduction modified mantle. Murphy et al. (2011) inferred from Nd isotope systematics that the parent magma was derived from a subcontinental lithospheric mantle.

Thermal regime of the ENA sub-province

Primary melt compositions and mantle potential temperature estimates (T P) were calculated for the North Mountain basalt using PRIMELT3 (Table 2). Previous T P calculations of the ENA basalts relied on the older versions of PRIMELT; however, there are significant improvements to the software including correcting better melt fractions, identifying the residuum mineralogy, and improved uncertainty in the thermal estimates (Herzberg and Asimow 2015). The dry peridotite source MgO contents were set at 38.12 wt%, whereas the bulk FeOt of the source was set to the lowest possible values (8.57–8.69 wt%) that would produce meaningful (no augite fractionation warning) results. The modeled FeO was calculated by setting Fe2O3 = 0.5 × TiO2 to reflect a reducing relative (FMQ < 0) oxidation state (Fig. 4d). The accumulated fractional melt (AFM) results from PRIMELT3 are plotted on a series of FeOt versus MgO diagrams that show the primary melt composition and the equilibrium melting olivine control line (Fig. 8a, b). The solidus, melt fraction, and pressure lines shown in the models are derived from Herzberg and O’Hara (2002), Herzberg et al. (2007), and Herzberg and Asimow (2008, 2015). The primary melt compositions and olivine control lines for the North Mountain basalt are shown in Fig. 8a. According to the classification of Le Bas (2000), the calculated primary magma compositions are picritic (MgO = 14.9–16.6 wt%) and experienced ~11.5 to ~20.9% olivine loss (Table 2). The calculated initial olivine has a forsterite value of 90 and the melt fraction (amount of melting from the source) for all samples falls within a narrow (0.30–0.32) range (Fig. 8b). The eruptive temperatures (T) and mantle potential temperatures (T P) are estimated to be 1330–1370 and 1430–1480 °C, respectively.

Table 2 Primary melt compositions and mantle potential temperatures of North Mountain
Full size table
Fig. 8
figure 8

a FeOt versus MgO of calculated primary melt compositions of the North Mountain basalt. The primary melt composition (solid circles) calculation results using PRIMELT3 (Herzberg and Asimow 2015). The small white circles represent the olivine control lines generated by the addition or subtraction of calculated equilibrium olivine compositions of the NMB. Dashed lines are melt fraction contours and the crosses indicate uncertainties in calculated FeO and MgO content related to ±1σ K Ol/LD FeO/MgO. b The primary melt compositions are plotted relative to the equilibrium olivine composition. The numbers (80, 85, 90, 95) represent Mg# (forsterite content) of olivine

Full size image

The mantle potential temperature estimates of the North Mountain basalt (T P = 1430–1480 °C) are within the range reported by Hole (2015) and Callegaro et al. (2013) for selected rocks from the Central Atlantic Magmatic Province (Virginia, Georgia, South Carolina, North Carolina and Iberia) in general (T P = 1450 ± 50 °C). A more thorough investigation of T P of basaltic rocks from Georgia, North Carolina, Pennsylvania, New Jersey, and Newfoundland shows that there is no significant difference (T Psouth = 1330–1490 °C; T Pnorth = 1390–1490 °C) along the length of ENA sub-province (Fig. 9a). Overall, the T P values of the CAMP are within or higher than ambient mantle (1300–1400 °C) but lower than temperatures expected from an anomalously (>1550 °C) hot mantle (Fig. 9a). Consequently, based on a limited dataset, Hole (2015) suggested that the T P values of the CAMP are due to mantle insulation by the continental lithosphere rather than a mantle plume, a viewed shared by Coltice et al. (2007), Callegaro et al. (2013), and Whalen et al. (2015). Furthermore, the similar T P values from South to North favor a thermal regime expected for a passive extensional (plate stress) model because a mantle-plume thermal regime is thought to change from the center to the margins (Herzberg and Gazel 2009). The fact that the basalts distal (~3000 km) from the inferred center of the CAMP have similar T P as the basalts more proximal (<1000 km) to the center implies a consistent ambient thermal regime in each basin.

Fig. 9
figure 9

a Mantle potential temperature (T P) estimates of basalts from the ENA sub-province relative to ambient mantle (1300–1400 °C) and hot mantle associated with mantle-plumes (>1550 °C). The average T P for the Southern and Northern Appalachians is 1440 °C. b Regional variability of the bulk FeOt content of the source required to produce a meaningful T P result. GA South Georgia basin, Georgia; NC Deep River basin (Durham sub-basin), North Carolina; PA = Gettysburg basin, Pennsylvania; PA-NJ West Newark basin, Pennsylvania-New Jersey; NS Fundy basin, Nova Scotia; NL Avalon Peninsula, Newfoundland (Chowns and Williams 1983; Grossman et al. 1991; Pe-Piper and Reynolds 2000; Herzberg and Asimow 2008)

Full size image

Compositional variability of the ENA mantle source

The accumulated fractional melting (AFM) primary melt compositions of this study show that the primary melt compositions from Northern Appalachians require an Fe-rich (FeOtaverage ≈ 8.6 wt %) mantle source and have a harzburgite residue (olivine + opx ± Cr-spinel), whereas basalt from the Southern Appalachians requires a less Fe-rich (FeOtaverage ≈ 8.3 wt%) mantle source composition where nearly two-thirds of the primitive melts have a spinel peridotite (olivine + opx + cpx + spinel-Al or Cr) residue and the rest indicate the primitive melts were in equilibrium with a garnet peridotite or harzburgite source (Fig. 9b). The implication is that the Northern basalts may be derived from an ‘Fe-rich’ mantle source and that the primary melts were in equilibrium with a harzburgitic residue although the original source may have been a spinel peridotite or a garnet peridotite (Herzberg and Asimow 2015). In contrast, southern basalts may be derived from a mantle source that was less ‘Fe-rich’ but that the residual lithology is typically spinel peridotite and thus the source was likely a spinel-bearing peridotite for most of the SENA basalt but in a few cases may have been a garnet-bearing peridotite.

Compositional variation of the ENA basalt

If the ENA basalts are related to mantle insulation (mantle global warming) and passive extension then the spatial-compositional variation observed from South to North must be related to inherent differences in the mantle sources that are beneath the Southern and Northern Appalachians (Weigand and Ragland 1970; Pegram 1990; Grossman et al. 1991; Puffer 2001; Marzoli et al. 2011; Murphy et al. 2011; Callegaro et al. 2013; Merle et al. 2013; Whalen et al. 2015). Most of the Triassic–Jurassic basins in the Northern Appalachians formed within the West Avalonia terrane, whereas the Triassic–Jurassic basins in Southern Appalachians (North Carolina to Florida) were formed within Carolinia or the Suwannee terrane (Dennis and Shervais 1996; Dennis and Wright 1997; Heatherington and Mueller 1999, 2003; Murphy and Nance 2002; Murphy et al. 2011; Callegaro et al. 2013). Crustal contamination and/or post-eruption hydrothermal alteration likely played a role during the ascent of the magmas through the crust before eruption but these processes are probably not responsible for the distinct regional chemical variation as they are highly localized and are unlikely to be uniform from region to region (Pegram 1990; Callegaro et al. 2013; Merle et al. 2013; Whalen et al. 2015).

The Sr–Nd–Pb isotopes across the ENA do not show systematic variation although the basaltic rocks from South Carolina and North Carolina have the lowest 87Sr/86Sri values (0.7045–0.7055), highest ε Nd(t) values (ε Nd(t) = + 1 to +4) and lowest Pb (207Pb/204Pb = 15.60–15.53; 208Pb/204Pb = 37.8–37.3; 206Pb/204Pb = 18.0–17.5) isotopic signatures (Merle et al. 2013; Callegaro et al. 2013). Callegaro et al. (2013) suggest that a maximum of ~10% recycled crust could explain the trace elemental and isotopic variability within the southern basalt and that the underlying mantle may be exceptional with respect to the rest of the CAMP. It is suggested that the subduction of upper continental crust (oceanic sediments) during the Paleozoic (Acadian Orogeny) and possibly delaminated lower crust were incorporated or reacted with the ambient shallow lhzerolitic mantle. Trace elemental modeling shows that the range of La/YbN ratio of the rocks from the Southern ENA can be derived by ~15 to ~25% partial melting of 5% spinel peridotite source with a starting composition similar to primitive mantle (Fig. 10). The moderately high amount of partial melting is supported by the generally higher Mg# (60–72) and Ni (88–390 ppm) concentration in the basalts from the Southern Appalachians. Moreover, the calculated melt fraction for the primary melt compositions is consistent with the trace element models and range from 6.3 to 28.2% with an average of 16.9% (Table S1). Although it is likely that the mantle source of the SENA basalt was affected by subduction-related processes (Pegram 1990; Heatherington and Mueller 1999, 2003; Callegaro et al. 2013), our trace elemental modeling indicates that the mantle source of the southern Appalachians may have had a trace element signature similar to primitive mantle.

Fig. 10
figure 10

The chondrite-normalized (La/Yb)N ratio versus TiO2 (wt%) of ENA basalts from the Deep River basin (NC), Gettysburg basin, (PA), West Newark basin (PA-NJ), Fundy basin (NS) and Avalon Peninsula (NL). The red and black model curves are based on partial melting of a 5% spinel bearing peridotite (olivine = 52%, opx = 25%, cpx = 18%) and a peridotite with 1% garnet and 4% spinel (olivine = 52%, opx = 25%, cpx = 18%). Starting composition is equal to primitive mantle values of McDonough and Sun (1995). The green and blue model curves are based on a partial melting of a 5% spinel bearing peridotite (olivine = 52%, opx = 25%, cpx = 18%) assuming different ‘mantle wedge’ compositions for the north and central Appalachians (Table 3). Batch melting equation: C L/C o = 1/D(1 − F) + F, C L concentration in the liquid, C o original rock composition, D bulk distribution coefficient, F weight fraction of melt produced. K d values: Ti, olivine = 0.04, opx = 0.1, cpx = 0.78, spinel = 0.048, garnet = 0.28; La, olivine = 0.007, opx = 0.0003, cpx = 0.056, spinel = 0.01, garnet = 0.001; Yb, olivine = 0.049, opx = 0.227, cpx = 0.28, spinel = 0.01, garnet = 8.5 (Arth 1976; Irving and Frey 1978; Fujimaki et al. 1984; Green et al. 1989, 2000; McKenzie and O’Nions 1991; Beattie 1994; Jenner et al. 1994; Johnson 1994; Salters and Longhi 1999; Klemme et al. 2006). The triangles are outliers from the Fundy Basin (NS)

Full size image

Compared to the southern ENA basalt, the basalt from the northern ENA appears to have a relatively complicated petrogenetic history (Whalen et al. 2015). A similar but slightly different interpretation is expressed for the northern basaltic rocks as it is thought that a greater number of accretionary events or metasomatic events by subducted sediments affected the sub-Avalonia lithospheric mantle during the Paleozoic and yielded different ‘subduction-related’ signatures (Pegram 1990; Murphy et al. 2011; Merle et al. 2013; Whalen et al. 2015). Therefore, the major and trace elemental signatures are probably representative of the source region rather than syn-magmatic or post-magmatic processes (e.g., crustal contamination). A spinel peridotite source with a primitive mantle composition is unlikely to be responsible for the (La/Yb)N ratios observed in the northern ENA basalts as the melt curves cannot reproduce the high (>3) values or high TiO2 content of the NMB or the Avalon dykes in Newfoundland (Fig. 10). The addition of a small amount of garnet to the mantle composition shifts the partial melting line upward and shows that lower amounts (5–15%) of partial melting can produce the compositional range of the basalts from the West Newark, Gettysburg, and Fundy basins and the dykes from the Avalon Peninsula but the PRIMELT3 calculations indicate that the primary melt compositions of the northern basalts were derived by high (~30%) melt fractions (Fig. 8a). The high melt fractions and low (≈10%) olivine loss in some of the models (NMB-8 and NMB-19A) would likely produce Ni-rich (>100 ppm) and high-MgO (>8 wt%) basalt. The MgO content for basalt that produced meaningful primary melt compositions typically have high MgO (>8 wt%) concentrations. However, the basalts from the northernmost (Nova Scotia and Newfoundland) ENA have lower Ni concentration (≤110 ppm) than the SENA (Ni = 88–390 ppm) but the basalts from Pennsylvania and New Jersey have Ni contents (Ni = 75–240 ppm) between the northernmost and southernmost. The ‘transitional’ composition of the basalts from the West Newark and Gettysburg basins also is reflected in their La/YbN ratios and TiO2 (wt%) content (Fig. 10).

The calculated melt fraction, trace element modeling, and bulk Ni and TiO2 contents are at odds suggesting that the mantle source of the NENA basalt must be different than the SENA. Mantle metasomatism is likely responsible for enrichment of some major and trace (K, Fe, Ca, Ti, Nb, and Ta) elements and isotopes that may be related to recent or even ancient events (Menzies et al. 1983; Dawson 2002; Ionov et al. 2002; O’Reilly and Griffin 2013). Mantle xenolith compositions reported from the Trans-Mexican Volcanic Belt (TMVB), considered indicative of a mantle-wedge source, show variability and trace element enrichment relative to primitive mantle (Mukasa et al. 2007). If a mantle xenolith (X-30 from Mukasa et al. 2007) composition from the TMVB is used as a proxy for the type of mantle that was present during the Early Jurassic beneath the North Appalachians then it appears that the high melt fractions and trace element modeling can be reconciled. The batch melting curve (green line) for the ‘mantle wedge’ peridotite is shown in Fig. 10. A small amount of melting (<10%) produces the high La/YbN ratios (>5) but larger amounts of melting (>15%) produce the range (La/YbN = 3–4) of observed values in the rocks from Newfoundland (Avalon Peninsula) and Nova Scotia (Fundy Basin). The TiO2 concentration was not reported by Mukasa et al. (2007) but we use a value of ~0.35 wt% which is high but lower than metasomatized spinel peridodite xenoliths (TiO2 ≤0.46 wt%) from southeastern Mongolia (McDonough 1990; Kononova et al., 2002). The model is in agreement with the calculated melt fraction (20–30%) but the mantle xenolith composition cannot, within reason, reproduce the lower La/YbN (<3) and TiO2 values of the basalt from the West Newark and Gettysburg Basins.

The basalts from the Mid-Atlantic States (New Jersey, Pennsylvania) are referred to as having a ‘transitional’ composition between the SENA and NENA basalts (Marzoli et al. 2011; Merle et al. 2013; Whalen et al. 2015). Although the ‘transitional’ basalts are still derived from a metasomatized mantle source, the amount of influence from subduction that the source experienced is thought to be different and contributed to their slightly different compositions as compared with the North Mountain basalt (Whalen et al. 2015). The differences in the structural controls related to subduction between the northern basalts and the transitional basalt could be the explanation for their different TiO2 and La/YbN ratios in Fig. 10 and why the ‘proxy mantle wedge’ source composition cannot reproduce the values from the West Newark and Gettysburg Basins. If we use a mantle source composition (Table 3) that is ‘transitional’ between SENA and NENA rocks then values similar to the range observed in the West Newark and Gettysburg Basins can be reproduced and consistent with the PRIMELT3 melt fraction calculation (>25% melting).

Table 3 Results of trace element modeling of the ENA flood basalts
Full size table

Variability in mantle source composition between the basalt of the SENA and NENA is largely attributed to differences in the precise processes that occurred during Paleozoic subduction and accretion of West Avalonia and Carolinia to eastern North American margin (Puffer 2001; Murphy et al. 2011; Callegaro et al. 2013; Merle et al. 2013; Whalen et al. 2015). However, there may be an additional reason for the chemical differences between the regions of the NENA. Wide spread Late Devonian silicic plutonism is a defining difference between the tectonic evolution of the Northern and Southern Appalachians (Murphy et al. 1999; Dorais and Paige 2000; Dorais 2003; Tomascak et al. 2005; Shellnutt and Dostal 2015). The plutonism is spatially restricted to the regions north of New Hampshire and Vermont but is concentrated in Maine, New Brunswick, and Nova Scotia (Murphy et al. 1999). Shortly after the emplacement of the batholiths, there was a significant eruption of TiO2-rich (>1.5 wt%) flood basalts within the Maritimes Basin during the Early Carboniferous (La Flèche et al. 1998; Dessureau et al. 2000). It is thought that the flood basalt province was related to a mantle plume that migrated across the region (Murphy et al. 1999; Dessureau et al. 2000). It is possible that the mantle plume contributed (via mafic magma injection) to regional scale crustal melting prior to the eruption of the flood basalts and produced the silicic batholiths (Murphy et al. 1999; Shellnutt and Dostal 2015). We suggest that additional enrichment of the NENA basalts may, in part, be related to mixing with melts derived from underplated mafic rocks associated with the Maritimes Basin basalts and/or the melt-extracted Maritimes Basin mantle source. The Early Carboniferous magmatism, whether mantle plume-derived or not, was spatially restricted to the Northern Appalachians and did not affect regions south of New England (New Hampshire/Vermont) as there is no evidence of Late Devonian magmatism in Pennsylvania, New Jersey, Maryland or Virginia.

Tectonic synthesis

In this section, we present a synthesis of the Paleozoic regional tectonic and structural evolution of Eastern North America that summarizes the factors that likely contributed to the modification of the mantle sources beneath the Northern (A–A′) and Southern (B–B′) Appalachians (Fig. 11). The two profiles are depicted in a series of tectonic evolution diagrams that span from ~450 to ~200 Ma and outline the major differences in the tectonic events of each region (Fig. 12).

Fig. 11
figure 11

Topographic map showing major lithotectonic provinces and boundary fault systems of the Appalachian orogeny. Modified from Hibbard et al. (2003, 2007) and van Staal and Barr (2012). The detailed evolution of the two profiles (A–A′ and B–B′) is shown in Fig. 12. Appal Structural front Appalachian Structural front, P. accretionary cplx post-accretionary complex

Full size image
Fig. 12
figure 12figure 12

a Sketch diagram illustrating the tectonic evolution of the Northern Appalachian orogeny along transect AA’ over the Salinic (442–425 Ma), Acadian (419–400 Ma) and Neoacadian (380–370 Ma) orogeny. Modified from van Staal and Barr (2012) and Hibbard et al. (2007). b Sketch diagram illustrating the tectonic evolution of the Central and Southern Appalachian orogeny along transect BB′ over the Taconic (455–443 Ma), Acadian–Neoacadian (419–400 Ma), Late Missisippian–Early Pennsylvanian (325–300 Ma), and Alleghenian (300–260 Ma) orogeny. Modified from Hatcher (1978), Murphy et al. (2010), and Tremblay and Pinet (2016)

Full size image

Structurally, the Northern Appalachians vary markedly from the southern and central Appalachians both in tectonostratigraphic component and deformation sequences (Fig. 12a, b). The difference is attributed to the shape of the original Laurentian continental margin, the shape of the Gondwana indenter, and the timing and nature of accretionary processes of the accreted terranes composed of recycled Grenville basement, Laurentian metasedimentary oceanic and arc-affinity material (Keller and Hatcher 1999). A series of episodic compressional events following the rifting of Rodinia began in the Ordovician and spanned much of the Palaeozoic era. The Taconic orogeny (Middle Ordovician—about 460 million years ago) resulted from the convergence of the Theia Sea, where earlier formed island arcs were obducted onto the Laurentian continental margin (Hatcher and Odom 1980; Faill 1997a, b, c). It consists of a complex series of orogenic events varying in intensity through time and along the orogen in eastern North America. The evidence for this orogeny is most pronounced in the northern Appalachians, but its affects stretch south to Tennessee in the Valley and Ridge province and Georgia in the Piedmont province (Glover et al. 1983; Dalziel et al. 1994; Faill 1997a). The Acadian Orogeny followed the Taconian (Early to Late Devonian—408–360 million years) and resulted from collision of the north-eastern portion of the North American Plate (Laurentia) with Western Europe affecting the north Appalachians from New York to Newfoundland (Rodgers 1987; Dalziel et al. 1994; Faill 1997b). The subsequent Alleghanian Orogeny (Late Carboniferous to early Triassic—300–250 million years ago) resulted from collision of the central and southern Laurentian continental margin with West Gondwana (Bradley 1989; Dalziel et al. 1994). This collision led to the formation of the Pangea supercontinent, with d’ecollement tectonism mostly in the central and southern Appalachians (Faill 1997c). Late Palaeozoic folding and igneous intrusion along the east coast of New England and parts of Atlantic Canada occurred (Bradley 1989). The age of these orogenies decreases eastward across the orogenic belt, suggesting that there was progressive eastward addition of arcs and continental fragments to the continental margin of North America (Zen 1968).

Widespread regional metamorphism and ductile deformation resulting from the Taconic orogeny were observed throughout the piedmont of the central and southern Appalachians. Radiometric age determinations (U/Pb zircon ages) indicate two major thermal events within the Baltimore Gneiss, north-eastern Maryland. The first event ranged from 1000 to 1200 Ma and occurred during the Greenville Orogeny. The second event ranged from 420 to 450 Ma (Tilton et al. 1970; Grauert 1973a, b, 1974; Muller and Chapin 1984; Stamatakos et al. 1996; Aleinikoff et al. 2002) and occurred during the Taconic Orogeny. Rb–Sr mineral ages (425–470 Ma) from a relatively undeformed pegmatite cutting the regional Glenarm schistosity after the formation of early folds (Glover et al. 1983) suggest that the peak of amphibolite facies metamorphism occurred from late Middle to early Late Ordovician time. Similar U/Pb zircon ages (489 Ma—crystallization age; 453 Ma—metamorphic age) were also recorded from the juxtaposed Baltimore mafic complex within the Glenarm series (Sinha et al. 1997). Localized retrograde metamorphism after the Ordovician thermal peak affects the Glenarm series to varying degrees (Wetherill et al. 1966, 1968; Muth et al. 1979). Young Rb/Sr biotite ages (290 Ma; Wetherill et al. 1968) reveal that the temperatures remained high enough for strontium diffusion after the Ordovician thermal peak until almost the late Pennsylvanian. This suggests that metamorphism extended from the Taconian to the Acadian Orogenies, accompanied deformation and produced widespread garnet-, staurolite-, kyanite-, and fibrolitic sillimanite bearing assemblages in the Glenarm series (Tilton et al. 1970; Grauert 1973a, b; Muller and Chapin 1984). Alleghanian cooling ages associated with later faulting are also suggested (Lang 1990; Muller and Chapin 1984). The deformation age is progressively younger to the east for the northern Appalachians yet the central and southern Appalachians progressively youngs westward (Hatcher and Odom 1980).

Post-Alleghanian–Variscan orogeny, eastern North America became a passive margin during the early Mesozoic as the Atlantic Ocean opened and Pangaea began to break-up. The extensional stress setting reactivated Paleozoic structural weak zones (e.g., Cobequid fault zone in Nova Scotia) that became the North American rift system, with rift basins and widespread mafic volcanism (Geoffroy 2005; Whithjack et al. 2012; Melankholina and Sushchevskaya 2015). Whithjack et al. (2012) separated the Jurassic CAMP-related magmatic activity along the eastern North American rift system into three regions of North, Central, and South, which correlate to our north, central, and south segments of the Appalachians. The north segments are composed of NE-SW Avalon dyke and lava flow sequences within Newfoundland and Jeanne d’Arc basin. Less intense magmatism is noted for the central and southern segment, but are composed of wide spread N to NE striking dyke, thick lava flow sequences, and intrusive sheets. However, no lava flow sequences are noted for the southern segment.

Melankholina and Sushchevskaya (2015) pointed out that there is a dissimilar isotopic composition among the north verses central segments magmatism. We consider the compositional differences between the Southern and Northern Appalachian basalts are probably related to the specific tectonomagmatic evolution of the accreted terranes (Carolinia and West Avalonia) as they each rifted from Gondwana and accreted to Laurentia during the Neoproterozoic to Late Palaeozoic (Murphy and Nance 2002; Nance et al. 2002; Whalen et al. 2015). Moreover, Late Devonian–Early Carboniferous magmatism in New England and Atlantic Canada was an additional influence of Middle Mesozoic mantle melting in the north (Fig. 12a). As tensional plate stresses acted on Pangaea during the Early Jurassic, rifting and basin formation was initiated followed by mantle melting and the eruption of the ENA basalts and emplacement of mafic dykes (Dostal and Durning 1998; Pe-Piper and Reynolds 2000). It is likely that the Southern basalts were derived by partial melting of a mantle source beneath Carolinia that was less affected by subduction and accretion, whereas the Northern basalts were derived by partial melting of mantle beneath West Avalonia that experienced melting associated with subduction and upwelling mantle and rifting during the Late Devonain to Early Carboniferous (Murphy et al. 2011; Callegaro et al. 2013; Merle et al. 2013; Whalen et al. 2015).

Conclusions

Basaltic rocks from the Early Jurassic ENA sub-province of the CAMP were likely produced under a thermal regime that was between ambient mantle conditions (T P ≈ 1440 °C) and hot mantle. The thermal conditions are consistent with melting above a passive extensional tectonic setting rather than an active mantle plume setting. The rocks from the Northern Appalachians were likely derived from an Fe-rich (FeOt = 8.6 wt%) ‘mantle wedge-type’ source that was affected by a Late Devonian–Early Carboniferous mantle plume. In contrast, the rocks from the Southern Appalachians were derived primarily from spinel peridotite that had less FeOt (FeOt = 8.3 wt%) and was not affected by Late Devonian-Early Carboniferous magmatism. The spatial-compositional variation of the ENA basaltic rocks is attributed to differing amounts of melting from mantle sources that experienced regionally unique subduction-related mantle enrichment prior to and during the accretion of Gondwanan terranes (Carolinia and West Avalonia) to eastern North America.