Continental crust forming the Newfoundland and Iberia basement consists of Precambrian and Paleozoic rocks accreted during Paleozoic closure of the Iapetus and Rheic oceans. The Taconian, Salinic, Acadian, and Alleghenian orogenies (Early to Middle Ordovician, early Silurian, Silurian–Devonian, and Carboniferous–Permian, respectively) resulted in accretion of a series of northeast-southwest–oriented terranes on the present Canadian Atlantic margin (van Staal et al., 1998; Waldron and van Staal, 2001; Percival et al., 2004). Iberia basement accreted during the mid-Devonian to Carboniferous (i.e., in the early part of the Carboniferous–Permian Hercynian [Variscan] orogeny). Easternmost Newfoundland and most of the Grand Banks platform consist of Avalon Terrane (Silurian–Devonian), although a sliver of Meguma Terrane (Carboniferous–Permian) is present at the southern edge of the Grand Banks, separated from the Avalon by a prominent magnetic anomaly (Collector anomaly) (Fig. F1).
The accreted terranes in Iberia are oriented northwest–southeast, but offshore they may have curved northeast and east as part of the Ibero-Armorican arc (Ziegler, 1982; Capdevila and Mougenot, 1988; Silva et al., 2000). Tonalites were intruded into the continental crust of the southern Galicia block (ODP Site 1067) in the late Proterozoic and were later deformed during the Hercynian orogeny (Table T1) (Rubenach, 1999). Permian-age zircons from a surrounding amphibolite also indicate late Hercynian magmatism at this site (Manatschal et al., 2001). Large granitoid batholiths intruded the onshore terranes during and following the Hercynian (Pinheiro et al., 1996). Granulites sampled from the northwest Galicia margin appear to have formed in the lower to middle crust, but they equilibrated at upper-crustal conditions in Late Carboniferous to Permian time (Table T1) (Fuegenschuh et al., 1998); this suggests that they may have been exhumed by a late phase of extension toward the end of the Hercynian orogeny.
Large-scale, original structural fabrics within the Paleozoic terranes are well defined by gravity and magnetic anomalies (Welsink et al., 1989; Srivastava et al., 1990a; Silva et al., 2000). A general concordance between these fabrics and subsequent Mesozoic rift patterns suggests that the Paleozoic structures exerted significant control on rift development (Wilson, 1988; Welsink et al., 1989; Pinheiro et al., 1996). Mesozoic rifting occurred during two phases, separated by a period of epeirogenic subsidence, as described below.
In Late Triassic to earliest Jurassic time, extension operated in a wide-rift mode and created large rift basins over a distance of at least 500 km across the Grand Banks-Iberia platform. The Jeanne d'Arc, Whale, Horseshoe, Carson, Salar-Bonnition, and Flemish Pass basins were formed in the present Grand Banks, and the Lusitanian, Porto, and possibly Galicia Interior basins opened on the Iberia margin (Fig. F1) (Wilson, 1988; Murillas et al., 1990). These basins accumulated red siliciclastic sediments during the Late Triassic (Carnian–Norian) and then were filled with widespread evaporite deposits during the earliest Jurassic (Hettangian–Sinemurian) (Jansa and Wade, 1975; Wilson, 1988; Rasmussen et al., 1998; Alves et al., 2003). This rifting was synchronous with rifting between Africa and North America farther south, but it did not lead directly to opening of an ocean basin between Newfoundland and Iberia.
The rifting was associated with relatively minor magmatic activity during the Carnian–Hettangian (Table T1). Most of the magmatism appears to have been limited to Iberia, where intrusions and flows affected the southwestern Iberia margin and the Central Iberian Zone (see Pinheiro et al., 1996, for a summary). At the southern edge of the rift, offshore Iberia, a Sinemurian diorite from Gorringe Bank (Carpena, 1984) probably was emplaced during this rift phase. There is only one occurrence of magmatism documented on the Newfoundland side, consisting of diabase intrusions in the Avalon Peninsula (Hodych and Hayatsu, 1980).
During the remainder of the Early Jurassic and through the Middle Jurassic, most of the Grand Banks-Iberia platform experienced epeirogenic subsidence without significant extension (Tankard and Welsink, 1987). An exception was in the Lusitanian Basin south of the Nazaré fault where an episode of extension and subsidence allowed deposition of a thick sequence of Sinemurian–Callovian carbonates (Rasmussen et al., 1998).
A second phase of rifting that began in the Middle Jurassic became more intense in the Late Jurassic and finally led to continental breakup and seafloor spreading late in the Early Cretaceous (Tucholke et al., 2007; Peron-Pinvidic et al., 2007). During this period, patterns of extension shifted from a wide-rift mode and eventually focused at future distal margins where continental crust separated. Based on analysis of seismic sequences and the timing and patterns of faulting on both sides of the rift, Tucholke et al. (2007) proposed three primary episodes of extension during this rift phase. The first began in the Middle to Late Jurassic and culminated near the end of Berriasian time, the second occurred during the Valanginian to Hauterivian, and the third ended near the Aptian/Albian boundary (Fig. F4). Each of these is discussed below.
In this paper, we use the timescale of (Gradstein et al., 2004), which the Leg 210 Scientific Party decided to standardize on for the present Scientific Results volume. We point out that this timescale is different from the Gradstein et al. (1995) and Channell et al. (1995) timescales used in the Leg 210 Initial Reports volume and also used by Tucholke et al. (2007). In particular, the bases of the Aptian and older stages in the Early Cretaceous are some 3–4 m.y. older in the Gradstein et al. (2004) timescale. This changes the timing of some events discussed by Tucholke et al. (2007); thus, some interpretations in the present text may differ from those in Tucholke et al. (2007).
Although the Middle Jurassic to Berriasian was a period of general extension across the Grand Banks-Iberia platform, the location, timing, and intensity of rifting varied considerably. In the Jeanne d'Arc Basin on the central Grand Banks, rifting is documented from late in the Middle Jurassic (late Callovian) into the Aptian, but it was most intense during the late Kimmeridgian to early Valanginian (Tankard et al., 1989). Along the southern Grand Banks in the Whale Basin there appears to have been weak Early to early Middle Jurassic extension, followed by increased rifting in the Bajocian–early Bathonian to late Kimmeridgian and then by limited extension through the late Tithonian (Balkwill and Legall, 1989).
On the present Iberia margin, rifting in the Lusitanian Basin occurred primarily in the Late Jurassic (Oxfordian to early Kimmeridgian), with minor extension continuing through the Tithonian to Barremian (Wilson, 1988; Wilson et al., 1989; Rasmussen et al., 1998). Along strike to the north, interpreted Tithonian to Berriasian shallow-water carbonates in the Galicia Interior Basin are highly faulted on seismic reflection sections (Murillas et al., 1990). On thinned continental crust at the deep western margin of Galicia Bank, faulted, shallow-water Tithonian to (?)Berriasian carbonates are succeeded by deeper-water carbonates and clastics (Boillot, Winterer, Meyer, et al., 1987). Deeper-water Tithonian claystones are present over crust along the southern margin of the Galicia continental block (Whitmarsh and Sawyer, 1996; Whitmarsh and Wallace, 2001). Thus major extension appears to have affected both the Galicia Interior Basin and the western and southern Galicia margins during the Tithonian and Berriasian.
Even though the above data for the two margins indicate that strong extension affected the northern part of the rift in the Tithonian to Berriasian, they seem to suggest little coeval extension in the southern rift. However, no precise information is available about timing and intensity of rifting in the Salar-Bonnition Basin or on the conjugate southern Iberia margin seaward and south of the Lusitanian Basin (Fig. F1). Thus, it is possible that this was a period of major extension in the central to southern part of the rift. As we discuss below, this extension probably focused at the future distal margins and resulted in separation of continental crust in the southern rift.
Our current understanding of the seaward limits of thinned continental crust along the Newfoundland and Iberia margins is shown in Figure F1. In the central Newfoundland Basin and under the Iberia Abyssal Plain along the southern margin of the Galicia block, these limits are reasonably well defined by seismic refraction results and by drilling. On the Newfoundland margin, SCREECH Line 3 refraction data suggest a 60-km-wide zone of very thin (~4 km) continental crust followed seaward by an 80-km-wide zone of exhumed, serpentinized peridotite (Lau et al., 2006b). Similarly, on SCREECH Line 2 about 60 km of thinned continental crust is succeeded seaward by exhumed mantle, with the transition probably occurring in the vicinity of Site 1276 (Van Avendonk et al., 2006). On the conjugate Iberia margin, a change from thinned continental crust to exhumed mantle is well defined by drilling results and by refraction data along CAM lines and the IAM9 line (Chian et al., 1999; Discovery 215 Working Group, 1998; Dean et al., 2000).
South of the Newfoundland and Tore seamounts, however, the crust-mantle transition is poorly constrained. On the Newfoundland margin, diapiric evaporites, most likely of Hettangian–Sinemurian age, in the Salar-Bonnition Basin (Austin et al., 1989) indicate that the seaward hinge line of this basin probably marks the minimum seaward extent of thinned continental crust. A single refraction profile in this area (Reid, 1994) suggests thin "crust" of uncertain (but possibly serpentinite) composition overlying serpentinized mantle seaward of the hinge line, consistent with a crust/mantle boundary near the hinge line (Fig. F1). Under the conjugate central Tagus Abyssal Plain, similar velocity structure may also indicate the presence of exhumed mantle, although Pinheiro et al. (1992) interpreted the basement as oceanic. The edge of thinned continental crust to the east is poorly defined; Pinheiro et al. (1992) suggested that it may be in the eastern part of their refraction line and within the limits of a refraction line described by Purdy (1975) to the south, although it could lie farther east.
Assuming that the limits of thinned continental crust shown in Figure F1 for the southern part of the rift are approximately correct, we can estimate the age of the boundary there based on arguments from seafloor magnetic anomalies. Srivastava et al. (2000) modeled magnetic anomalies within the rift and proposed identifications of the M-series reversals. There are some significant discrepancies between their identifications of older anomalies and the limits of thinned continental crust shown in Figure F1. For example, they identify anomalies M17 and M20 over probable continental crust beneath the Salar-Bonnition Basin and over the thin continental crust interpreted along SCREECH Line 3. In addition, most of the M-series anomalies appear to be over exhumed mantle, so there are major questions about the source of the magnetic anomalies. The source and identification of the anomalies, as well the extension rates that they imply, are discussed in detail in "Magnetic Anomalies in Zones of Transitional Extension: Character, Origin, and Implications." In that discussion, we conclude that there is large uncertainty about the identification of anomalies older than ~M8, although the implied extension rates are reasonable. If we take the magnetic anomaly identifications of Srivastava et al. (2000) at face value, the edge of thinned continental crust on the Newfoundland margin falls at about anomaly M16 (~141 Ma; late Berriasian), and on the Iberia side it is at about anomaly M17 (~143 Ma; middle Berriasian). This suggests that extension of continental crust in the southern part of the rift probably proceeded through Tithonian to Berriasian time and culminated near the end of the Berriasian.
The timing of extension discussed above is consistent with ages of known magmatic activity, metamorphism, and crustal cooling apparently associated with rifting (Table T1). Hornblende ages in amphibolites at ODP Site 1067 and 1068 on the southern Galicia margin in the central part of the rift suggest rifting and cooling of these rocks at ages as early as ~167 Ma (Bathonian) to ~153 Ma (Kimmeridgian) and continuing to ~132 Ma (Hauterivian) (Table T1). Igneous rocks in the younger part of this time range (<145 Ma; Berriasian–Hauterivian) are scattered through the central to southern part of the rift, most notably in Gorringe Bank, the Lusitanian Basin, and the Whale Basin; however, igneous rocks of this age are uncommon in the northern part of the rift. At the southern edge of the Galicia block (ODP Site 900), a 40Ar/39Ar plagioclase age of 136.4 mark retrograde metamorphism of gabbros as they were exhumed by rifting (Féraud et al., 1996).
From all the above observations, we conclude that scattered but locally strong rifting extended the continental crust in the Newfoundland-Iberia rift from as early as Middle Jurassic time (Bajocian–Bathonian) to the Berriasian and Valanginian. Rifting was most intense in the southern to central part of the rift during the Middle Jurassic to middle Late Jurassic, followed by strong extension that probably affected the entire length of the rift in Tithonian–Berriasian to early Valanginian time. The later stage of rifting caused widespread extension of continental crust across the Galicia block in the northern rift, but in the southern half of the rift it appears to have culminated in separation of the crust (Figs. F1, F4).
In the northern, Flemish Cap-Galicia Bank segment of the rift during this period, only minor extension occurred in the Galicia Interior Basin. Interpreted Valanginian–Hauterivian sedimentary sequences there show some faulting, primarily toward the center of the basin, but the faults are scattered throughout the basin and most have small offsets (Murillas et al., 1990; Pérez-Gussinyé et al., 2003). The southern Galicia margin also shows signs of fault slip along a few major fault blocks in continental crust, whereas other blocks were inactive (Tucholke et al., 2007). In conjugate crust of the Grand Banks, intense fault-controlled subsidence was sharply reduced in the Jeanne d'Arc Basin compared to the preceding rift episode, and basin deposition changed from restricted marine or nonmarine to more open-marine conditions (Tankard et al., 1989).
In sharp contrast to the above, strong extension occurred along the western margin of Galicia Bank and necessarily along the conjugate seaward edge of Flemish Cap. Valanginian–Hauterivian sedimentary sequences on western Galicia Bank were faulted and rotated, primarily toward the end of this period (Fig. F5) (Mauffret and Montadert, 1987; Tucholke et al., 2007). Fission-track ages of ~129–126 Ma (Barremian) on apatites in granulites sampled from northwest Galicia Bank are interpreted to record the extensional exhumation of continental rocks on this margin (Table T1) (Fuegenschuh et al., 1998).
At the latitude of the ODP Leg 103 drilling transect (Fig. F1), the boundary between the western, feather edge of thin continental crust and exhumed mantle lies between Site 640 and Site 637 (drilled on a peridotite ridge), ~50–60 km east of anomaly M0 as identified by Srivastava et al. (2000). At an extension half-rate of 9–13 mm/yr (see "Implied Extension Rates"), this boundary would date to ~129–131 Ma (Hauterivian/Barremian boundary). Farther south along the southwest margin of the Galicia continental crust, the crust/mantle boundary is about the same age. ODP Sites 899, 897, and 1070 were drilled on peridotite ridges approximately on anomalies M5 (130 Ma), M3 (129 Ma), and M1 (127 Ma), respectively, whereas Site 1069 penetrated Tithonian and older (?Paleozoic) sediments over presumed continental crust west of peridotite basement cored at Site 1068 (Fig. F1) (Sawyer, Whitmarsh, Klaus, et al., 1994; Whitmarsh, Beslier, Wallace, et al., 1998). Depending on where the boundary lies between Sites 899 and 1069, it could predate 130 Ma (late Hauterivian) by as much as several million years. Taken together, the above age data suggest that separation of continental crust between Galicia Bank and Flemish Cap occurred from near the Hauterivian/Barremian boundary into the early Barremian. This pattern is consistent with the time at which the last major normal faulting occurred in continental crust along the west-central margin of Galicia Bank (Fig. F5).
During the Valanginian–Hauterivian little to no extension occurred in the continental crust of the southern Grand Banks and the southern Iberia margin (Balkwill and Legall, 1989; Wilson, 1988; Wilson et al., 1989; Rasmussen et al., 1998). Instead, extension concentrated within the deep basin, starting from the southern limit of continental crust on Galicia Bank and reaching to the southern edge of the rift. It appears that mantle was being exhumed in that area, and we consider this period, up to near the time of Chron M3 (~128 Ma), to be one of transitional extension (TE1) (Fig. F1). Normal seafloor spreading was not yet established at this time.
Evidence for mantle exhumation during this period comes from both drilling results and interpretations of seismic refraction data. Drilling at Sites 897, 899, and 1068 (Fig. F1) established that basement in those locations consists of serpentinized peridotite with very minor amounts of igneous rocks (Sawyer, Whitmarsh, Klaus, et al., 1994; Whitmarsh, Beslier, Wallace, et al., 1998). A large set of seismic refraction and reflection studies around the southwest Galicia margin (e.g., CAM, IAM9 in Fig. F1) documents thin "crust" (2–4 km) that has low seismic velocities of ~4–4.5 km/s (Chian et al., 1999; Whitmarsh et al., 1990; Discovery 215 Working Group, 1998; Dean et al., 2000). On the basis of seismic velocities, velocity gradients, and the general absence of a Mohorovicic (Moho) reflection, this basement has been interpreted as intensely serpentinized mantle peridotite, with serpentinization progressively reduced with depth (~7.2 km/s increasing to ~7.9 km/s) (Discovery 215 Working Group, 1998; Chian et al., 1999). The upper basement layer on the IAM9 reflection profile published by Pickup et al. (1996) also has a markedly nonreflective character that they interpreted to be the result of intense serpentinization. From refraction data on conjugate SCREECH Line 2 across the Newfoundland margin, Van Avendonk et al. (2006) interpreted a narrow zone of basement between Sites 1276 and 1277 as serpentinized mantle (Fig. F1).
Farther south in the Newfoundland Basin, velocity modeling along SCREECH Line 3 (Fig. F1) suggests an ~80-km-wide zone of exhumed, serpentinized mantle reaching seaward from highly thinned continental crust to about anomaly M3 (Lau et al., 2006b). Three other refraction studies in the conjugate TE1 zones (SR1, R94, and P92; Fig. F1) show a thin (1–3 km), low-velocity "crust" overlying a ~7.1–7.9 km/s layer that is 2–6 km thick and is presumed to be partially serpentinized mantle. The similarity of this velocity structure to that of serpentinized basement along the CAM and IAM9 profiles suggests that TE1 basement in the southern part of the rift is also serpentinized mantle.
As previously noted, Srivastava et al. (2000) proposed identifications of M-series magnetic anomalies within the rift. These anomalies span the exhumed mantle in TE1 and also reach into the older part of TE2, discussed in the next section. The source and identification of the anomalies, and the implied extension rates, are discussed in "Magnetic Anomalies in Zones of Transitional Extension: Character, Origin, and Implications."
Separation of Flemish Cap from Galicia Bank appears to have been completed by about early Barremian time, with mantle subsequently exhumed between these continental edifices (see discussion of peridotite ridges, below). Although separation of continental crust is often considered to mark the cessation of continental rifting and the coincident start of seafloor spreading (e.g., Falvey, 1974), there is good evidence in the Newfoundland-Iberia rift that significant intraplate extension continued well away from the plate boundary throughout Barremian and Aptian time (i.e., from about Chron M3 to younger than M0). Furthermore, a major seismic stratigraphic marker in the sedimentary record on both margins dates to the Aptian/Albian boundary and in the past was interpreted as the "break-up unconformity," presumed to mark the onset of seafloor spreading. This marker is the "U reflection" in the Newfoundland Basin (Fig. F3) (Tucholke et al., 1989), the "orange reflection" around DSDP Site 398, and the lateral equivalent of the orange reflection around the Galicia margin (Figs. F5, F6) (Groupe Galice, 1979; Mauffret and Montadert, 1988) (see "Sedimentary Record in the Deep Basins"). These seismic stratigraphic markers are collectively referred to here as the "Aptian event." The Aptian event is the most prominent Mesozoic sedimentary horizon on either margin, and it records a major tectonic and/or sedimentary event. We consider the Barremian and Aptian up to this marker to constitute a second period of transitional extension (TE2) (Figs. F1, F4). Below, we summarize the evidence for intraplate extension during this period and offer a possible explanation for its cause and for the cause of the Aptian event.
Evidence for Barremian–Aptian extension in the Jeanne d'Arc Basin was presented by Driscoll et al. (1995), who related late Barremian–early Aptian and late Aptian unconformities developed there to rift-onset episodes. Magmatism that may have been related to extension is represented by Barremian to Aptian basaltic flows and sills on the southern Grand Banks and by Aptian lamprophyre dikes that intruded Paleozoic rocks in northern Newfoundland (Table T1).
On a broader scale, a major unconformity termed the Avalon unconformity is developed across the Grand Banks, and it appears to mark a major change in stress regime that may coincide with the Aptian event. The unconformity generally separates faulted and folded, pre-Albian sequences below from flat-lying Albian and younger sediments above. Grant et al. (1988) associated this unconformity with "a decline in stress regimes as the continental crust between Flemish Cap and Galicia Bank separated and seafloor spreading began in the late Aptian." On the conjugate margin, Wilson (1988) also discussed an Aptian unconformity in the Lusitanian Basin; he did not specifically relate it to an extensional episode, but he suggested that it might have developed at the onset of seafloor spreading.
In contrast to the above, the deep-basin parts of the rift show relatively little seismic stratigraphic evidence for faulting and tectonic rotation of Barremian–Aptian sedimentary sequences below the Aptian event. A few places in the Newfoundland Basin show this deformation (e.g., Fig. F3, just west of anomaly M1), but in most areas the Aptian event is a very high amplitude reflection that masks the deeper section, and any extensional structures that might be present there are difficult to resolve. Over the continental crust of Galicia Bank, the Barremian–Aptian sequence usually is relatively level between basin-bounding blocks (Fig. F5); where the sequence laps high onto bounding blocks its configuration appears to reflect sediment input from local sources rather than tectonic movement.
In the exhumed mantle south and west of Galicia Bank, age data on igneous intrusions and other geological data on faulting and uplift appear to present a different picture. At the peridotite ridge where ODP Site 1070 was drilled, ~127-Ma basement (anomaly M1) was intruded by gabbroic pegmatites at 127 123.9 peratures of ~500°C (Table T1). However, plagioclases in the samples with the latter two ages did not cool below their blocking temperature (~150°–250°C) until ~117–101 Ma. There are two possible explanations for these age offsets. One is that the gabbros were intruded at depth, had a long cooling history, and later were quenched as they were exhumed by local, late-stage faulting (Whitmarsh and Wallace, 2001). The other is that the gabbros were originally intruded near the basement surface and thus cooled quickly, but the plagioclases were recrystallized by continuing hydrothermal activity and their ages thus reflect the end of retrograde metamorphism (Jagoutz et al., submitted [N1]). Presently it is not possible to distinguish between these alternatives. Plagioclase ages of 115.7 other gabbro pegmatites intruded in the peridotite basement at Site 1070 (Table T1), and these could reflect ages of original magmatism or late-stage cooling as noted above. In any of these cases, the 117- to 101-Ma activity at Site 1070 would have occurred 10–26 m.y. after basement was first emplaced and ~90–340 km away from the plate boundary, assuming an extension rate of 9–13 mm/yr (see "Implied Extension Rates"). The ~124-Ma magmatism may have been associated with local extension, and it occurred ~30–40 km from the plate boundary.
There is some evidence that the nearby peridotite massifs drilled at Sites 897 and 899 (Fig. F1) were uplifted by late-stage normal faulting. As noted earlier, magnetic anomaly identifications indicate that basement was emplaced at ~129 Ma at Site 897 and ~130 Ma at Site 899. At these two sites, serpentinized peridotite olistostromes and debris flows covered the peridotite basement during early Aptian and early late Aptian time, respectively. The deposits have been interpreted as being derived either locally at the crests of then-existing basement highs (Gibson et al., 1996; Whitmarsh et al., 2001) or as being emplaced by long-distance flows in a basin-plain setting before normal faulting uplifted the massifs (Comas et al., 1996). If the Comas et al. (1996) interpretation is correct, the flows were uplifted no earlier than early late Aptian time (~120 Ma), some 9–10 m.y. after the basement was first emplaced and ~80–130 km from the plate boundary, assuming that the extension rate was 9–13 mm/yr. The occurrence of such intraplate extension would lend support to the idea that the age data at Site 1070, noted above, can be explained by late-stage normal faulting.
To the north, the peridotite ridge drilled at ODP Site 637 is ~40 km east of anomaly M0 as identified by Srivastava et al. (2000), and at an extension rate of 9–13 mm/yr it should date to ~128–129 Ma. High-temperature shearing of a mylonitized dikelet in peridotite sampled from this ridge ~90 km north of Site 637 predates 122.0 (Féraud et al., 1988), but plagioclase in the same sample cooled through its blocking temperature about 5 m.y. later at 117 al., 1989) (Table T1). As at Site 1070, the late plagioclase cooling might be due to late faulting or it might mark the end of hydrothermal retrograde metamorphism. In the same location, however, gabbros were emplaced in the peridotite ridge at ~122 Ma and then later sheared (Table T1) (Schärer et al., 1995, 2000), indicating that normal faulting occurred >6–7 m.y. after the basement was emplaced and at a distance at least ~50–90 km from the plate boundary.
A sequence of minor magmatic events, much like those observed at Site 1070 off Iberia, occurred on the Newfoundland margin at Site 1277. Peridotite basement there was emplaced at ~127 Ma and coincidentally was intruded by an alkaline gabbro dike (128 again at 113.2 T1) (Jagoutz et al., submitted [N1]). Gabbro cobbles overlying the peridotite have much younger 40Ar/39Ar plagioclase ages of ~92–69 Ma (Table T1). The 113-Ma intrusion may have been related to extension up to 130–180 km distant from the plate boundary (assuming 9–13 mm/yr extension rates). However, it is unclear whether the younger ages define original magmatic ages that could have been associated with intraplate extension or whether they mark the end of retrograde metamorphism associated with hydrothermal activity. Either is possible; as discussed in "Postrift Magmatism," the Newfoundland Basin contains clear evidence for postrift magmatism.
To explain the observed and inferred examples of intraplate extension noted above, Tucholke et al. (2007) proposed that the entire rift continued to experience elevated in-plane tensile stress throughout Barremian and Aptian time. They suggested that this occurred because strong, probably subcontinental mantle lithosphere was being exhumed, which strengthened the plate boundary and thus distributed strain across the rift. Intraplate extension most likely would have concentrated in weak, serpentinized areas, consistent with the above observations that normal faulting appears to have been more common in the exhumed mantle than in the continental crust of Galicia Bank. Tucholke et al. (2007) also proposed that the rising asthenosphere finally breached the probably subcontinental mantle lithosphere near the Aptian/Albian boundary. This released in-plane tensile stress and led to magmatic seafloor spreading. The plate-boundary release could have resulted in a pulse of relative in-plane compression, causing local and regional lithospheric deformation (Braun and Beaumont, 1989; Kooi and Cloetingh, 1992). If this deformation stimulated mass wasting and mobilized turbidity currents it could account for the reflective sedimentary layers that define the Aptian event in seismic reflection records (Figs. F3, F5, F6). According to this hypothesis, the Aptian event is not a "break-up unconformity" in the sense that it marked the separation of continental crust. Instead, it coincides with much later separation of the probably subcontinental lithospheric mantle.
The above interpretation implies that substantial amounts of mantle lithosphere were exhumed within the rift during TE2. As shown in Figure F1, peridotite basement was drilled in TE2 at both Sites 1070 and 1277, and basement in the CAM refraction data and part of the IAM9 line south and east of Site 1070 has been interpreted as serpentinized mantle. Refraction data at HU18, W96, and SR2 also suggest thin "crust" (possibly serpentinite) overlying serpentinized mantle, and the seaward end of SCREECH Line 2 has a velocity signature that may represent mixed serpentinized mantle and oceanic crust.
Elsewhere, the SCREECH Line 1 refraction data in TE2 indicate that very thin oceanic crust overlies a thick layer of serpentinized mantle. The seaward portions of SCREECH Line 3, IAM9, and W96 have been interpreted as normal-thickness ocean crust. However, the IAM9 line ends among seamounts north of the Tore seamount complex and the end of SCREECH Line 3 is near the Newfoundland seamounts, so it is possible that crustal characteristics at the seaward ends of these lines were affected by seamount volcanism (Tucholke et al., 2007). In addition, they may have been affected by magma emanating from a plume at the southern margin of the rift, as discussed in the next section. The overall picture is that there was significant mantle exhumation in TE2, but there appears to have been an increasing magmatic component with time that may have included at least local creation of oceanic crust.
Five drill sites (637, 897, 899, 1070, and 1277) have sampled basement at either the transition between TE1 and TE2 or farther seaward in TE2 (Fig. F1). The basement is dominantly exhumed mantle with a richly variable composition, and it includes websterites, harzburgites ( spinel), lherzolites (amounts of dunite. Igneous components are minor. Mafic rocks include basalt, diabase, and gabbro. Minor alkaline veins are present in the peridotites at Site 1277 (Müntener and Manatschal, 2006), and weakly alkaline rocks are also documented at Site 899 (Cornen et al., 1996b).
Most of the peridotites show broadly similar thermal and alteration histories, although there are local differences (e.g., Agrinier et al., 1988; Evans and Girardeau, 1988; Girardeau et al., 1988; Beslier et al., 1996; Cornen et al., 1996a). In general, these peridotites experienced high-temperature deformation at ~900°–1100°C and pressures of ~7–10 kbar, followed by multiphase metamorphism at decreasing temperatures in the presence of hot fluids. Samples were often mylonitized and then partially reequilibrated at temperatures below ~850°–700°C. At lower temperatures of ~50° to ~300° they were serpentinized in the presence of infiltrating seawater; calcite impregnated and/or filled veins in the brittlely deformed rocks at temperatures as low as near-seafloor conditions (Skelton and Valley, 2000). These features are consistent with the expected history if the mantle were exhumed along one or more major shear zones.
Serpentinized spinel harzburgites recovered at Site 1277 on the Newfoundland margin are among the most depleted abyssal peridotites observed worldwide, probably having experienced 14%–25% melting in the spinel peridotite field (Müntener and Manatschal, 2006). Intervals of normal mid-ocean-ridge basalt (N-MORB) were recovered above the peridotites (Robertson, this volume), but the combined total thickness of these rocks is small (~9 m), which is highly inconsistent with the degree of melting documented in the underlying mantle. Based on mineral compositions, Müntener and Manatschal (2006) proposed that the depleted peridotites are subarc mantle, from which melt was extracted probably during the Paleozoic closure of the Iapetus and Rheic oceans in this region. If this is correct, the Site 1277 peridotites were part of the mantle lithosphere beneath the prerift continental crust before they were exhumed in approximately late Barremian time. Their prior history of extensive melting precluded generation of much magma when the peridotites were exhumed.
Compared to the Newfoundland samples, spinel peridotites on the Iberia margin are more fertile but show a large variation in degree of melting, and locally they exhibit evidence of equilibration in the plagioclase stability field (Müntener and Manatschal, 2006). On the basis of whole-rock major and platinum-group element (PGE) geochemical data, Hébert et al. (2001) argued that the rocks at Site 1070 (and also to the east at Site 1068) represent subcontinental mantle. Chemical heterogeneity of the peridotites might be explained by local synrift melt percolation and equilibration in the plagioclase stability field (e.g., Hébert et al., 2001), by subsolidus deformation related to extension (Beslier et al., 1996), or by both these processes (Cornen et al., 1996a). In any case, the amount of melt generated appears to have been insufficient to produce significant intrusions or flows, minor occurrences of which have been sampled only at Sites 899 and 1070 in TE1 and TE2. Basalt and diabase clasts in breccias overlying peridotite basement in Hole 899B have transitional to enriched mid-ocean-ridge basalt (E-MORB) characteristics (Seifert and Brunotte, 1996). At Site 1070, a 2.7-m-thick gabbro pegmatite lies at the top of the peridotites; the isotopes and modeled bulk composition of this gabbro also are similar to E-MORB in most respects (Beard et al., 2002).
The rift axis at the southern edge of the Newfoundland-Iberia rift (Fig. F1) was affected by plume magmatism from about the time of Chron M4 (latest Hauterivian; near the transition from TE1 to TE2) until some time after Chron M0 (Aptian). This magmatism formed the Southeast Newfoundland Ridge and probably part of the conjugate Gorringe Bank. It also created volcanic ridges that extended both to the south (forming the J Anomaly Ridge and the conjugate Madeira-Tore Rise) and to the north (Tucholke and Ludwig, 1982; Tucholke et al., 1989). It is uncertain how far the melt from the plume may have reached northward into the rift, but the amplitude characteristics of the magnetic J Anomaly (see "Magnetic Anomalies in Zones of Transitional Extension: Character, Origin, and Implications," below) suggest that it could have had an effect all the way to the southern Galicia margin. This may be recorded at Site 899 where a plume contribution to the E-MORB basalts has been suggested by Seifert and Brunotte (1996). Such magmatism might help to explain the oceanic-type crust indicated by refraction data at the seaward ends of the SCREECH and IAM9 lines (Fig. F1).