Available evidence from Site 1277, as summarized in Figure F14, is limited, and thus several different tectonic models need to be considered. These models should also take account of evidence from the Iberia-Galicia margin and the Alps-Apennines. Debatable aspects include the source of the serpentinized peridotite and related gabbroic intrusive rocks; the setting of formation of the mylonitic fabrics and ductile-type deformation within both the basement serpentinized peridotite (Unit 2) and the clasts in the overlying sediments (Unit 1); the setting of basaltic volcanism and its magmatic source; and the later history of brittle deformation, fracturing, veining and hydrothermal alteration, and uplift of the peridotite ridge.
The serpentinized spinel harzburgite is chemically highly depleted and can be considered to resemble depleted mantle-type sequences of suprasubduction-type ophiolites (Müntener and Manatschal, 2006). One possibility is that the harzburgite records the lower part of an ophiolite that was emplaced during Paleozoic closure of the Iapetus Ocean, similar to the Ordovician Bay of Islands ophiolite (Newfoundland), or the Troodos ophiolite (Cyprus). In the southwest Pacific region, the Neogene–Holocene Woodlark Rift dissected a regionally emplaced meta-ophiolite, as confirmed by the presence of ophiolite-derived clastic debris on the footwall of the rift, drilled during ODP Leg 180 (Taylor, Huchon, Klaus, et al., 2000; Robertson et al., 2001). If the Site 1277 ultramafic rocks represent part of an unmetamorphosed ophiolite they would have been preserved at a high structural level, and thus no deep exhumation of mantle would be needed to exhume them during rifting of the Newfoundland margin. However, an origin solely as an emplaced ophiolite is unlikely in view of the width and length of the inferred zone of exhumed sub-continental mantle lithosphere on both sides of the Atlantic (as wide as 150–180 km wide) and the absence of recovery of any higher-level ophiolitic rocks (e.g., diabase).
A second possibility is that the serpentinized spinel harzburgite formed at an "unusual" seafloor spreading center. In such an interpretation it would, however, be necessary to explain why the ultramafic rocks at Site 1277 are so depleted and how these rocks were exhumed. One possibility to explain the depletion is that seawater gained access to the mantle from above, allowing hydrous melting and thus producing unusually depleted magmas. It is generally believed, however, that the chemical depletion of ophiolitic mantle harzburgite can only be explained by the introduction of water from below related to subduction (e.g., Pearce et al., 1984). In a seafloor spreading model the ultramafic rocks might have been exhumed as a result of slow spreading, creating exhumation structures similar to the megamullions of the Mid-Atlantic Ridge (Tucholke et al., 1998). In this interpretation, no crustal sequence (e.g., diabase/basalt) need have existed locally above the ultramafic rocks; however, large volumes of basalts would still be expected related to regional seafloor spreading, and these do not occur at Site 1277.
A third alternative, favored here, is that the recovered serpentinite represents mantle that is now chemically depleted because it experienced melt extraction in a subduction-related hydrous setting at some earlier time (i.e., prior to North Atlantic rifting) (Müntener and Manatschal, 2006). The likely setting would be subduction related to closure of the Iapetus or Rheic oceans during the Ordovician to Permian.
In this third interpretation, it is necessary to explain why the Iberia margin ultramafic rocks do not show a similar depletion. One possibility is simply that the Iberia source mantle was located far from the inferred subduction zone, for example mantle that was exhumed from beneath the edge of the African (Gondwana) craton. Another possibility is that the Iberia mantle was depleted like that at Site 1277, related to melt extraction in a subduction setting, and that the mantle was then "refertilized" at a later stage, either prior to or during Atlantic rifting. There is little evidence, however, to support this possibility at present.
In summary, an origin as subcrustal mantle that was affected by a prior subduction event is inferred. This implies that continental crust originally existed above the mantle at Site 1277 but was removed by extensional faulting. The mid-ocean-ridge–type basalts were presumably extracted from less depleted mantle that is in the vicinity of Site 1277 but has not been sampled.
Several alternatives can be considered for the inferred high-temperature ductile deformation and cataclasis within the peridotite basement (Unit 2) and also within many of the clasts in the overlying the mass flows (Unit 1).
A first possibility, favored here, is that the ductile deformation took place on one or more deep-seated extensional faults within a core-complex system and that any overlying continental crust was removed, allowing the exhumation of subcontinental mantle lithosphere. A second possibility is that the contact between Units 2 and 1 records detachment faulting between oceanic mantle and oceanic crust (e.g., related to "ultra-slow seafloor spreading"). It is difficult to distinguish between these two options based on the structural evidence alone because neither continental crust (e.g., schist and gneiss) nor normal oceanic crust detritus (e.g., lava and diabase) were observed within the debris flows at Site 1277. The implication is that the upper plate of whatever crustal type was detached from the area at a relatively early stage such that all the detritus at Site 1277 was derived from the exhumed lower plate.
A third possibility is that the observed deformation records normal faulting at a high structural level. This is unlikely in view of the high-temperature ductile deformation observed within the basement (Unit 2) and also in some clasts within the mass flows (e.g., serpentinite mylonite). Also, the cataclasis at the top of the basement (Unit 2) and the cataclastic fabrics in some clasts within the mass flows (Unit 1) resemble features related to the low-angle extensional faulting in the Alpine ophiolites, as summarized in the previous section.
Several factors are consistent with the presence of a regional-scale detachment at Site 1277, notably the absence of crustally derived (terrigenous) sediment and the fact that the mass flows (Unit 1) include clasts that appear to be have been eroded locally from the detachment, including fault gouge material. The primary mantle tectonite fabric in the basement cores is commonly steeply inclined and is itself cut by sub-vertical fractures and veins (Shipboard Scientific Party, 2004b) (e.g., Sample 210-1277A-9R-4, 26–46 cm). If it is assumed that the primary mantle fabric was originally gently inclined, as in many ophiolites (Nicolas, 1989), it is possible that the basement was rotated around a horizontal axis within an extensional fault system and then cut by later high-angle brittle fractures. However, the orientation of mantle fabrics in ophiolites is known to be variable. The fissures and veins in the mass flows are known to have developed subvertically (>80°) because they cut well-laminated, subhorizontal (<20°) sediments that preserve the paleohorizontal at the time of deformation. Similar subvertical fissures and veins cut the underlying serpentinized harzburgite, including the inferred high-temperature ductile shear zones. This suggests that any significant rotation was accomplished by exhumation before the serpentinite mass flows and basalts covered the seafloor.
Assuming that a regional-scale detachment was developed, two alternate end-member geometries can be considered. The first is an upward-concave fault (listric), as seismically imaged on the Galicia margin inboard of the locus of final breakup of continental crust (Reston et al., 1995; Krawczyk et al., 1996). The second is a downward-concave fault, which has not been directly observed in the Atlantic or elsewhere but which is inferred to be a geometric requirement of mantle exhumation (e.g., Tucholke and Lin, 1994; Lavier and Manatschal, 2006). A system of rolling hinge faults could involve near-simultaneous exhumation of both upper mantle and crust, potentially on both margins of the zone of continental breakup. A wide zone of thinned continental crust is now believed to be present along the Newfoundland margin, as suggested by interpretations of the SCREECH 2 and SCREECH 3 seismic reflection data (Van Avendonk et al., 2006; Lau et al., 2006). Similar crustal thinning was previously suggested for the Iberia margin (Reston et al., 1995; Krawczyk et al., 1996), as noted above; however, the actual settings of exhumation of subcontinental mantle and continental crust may be complex and regionally variable, involving faulting of different scale, geometry, and timing that cannot yet be constrained by drilling or geophysical evidence.
Two origins for these flows can be considered, taking account of evidence from the Iberia-Galicia margin and the Alps-Apennines. Both assume a relatively local origin for the mass flows.
First, the peridotite breccias may have been deposited during the exhumation of mantle in a continuous process (i.e., serpentinite mass flows are an inherent product of the process of exhumation). A possible problem here is that the exhuming mantle was not simultaneously covered by terrigenous sediment derived from the upper plate, whereas elsewhere topographic lows were flooded by terrigenous turbidites (e.g., Site 1276 and DSDP Site 398) (Tucholke, Sibuet, Klaus, et al., 2004) or accumulated a terrigenous component (Iberia margin Site 897) (Comas et al., 1996).
A second alternative is that the serpentinite breccias formed later than the exhumation of the mantle as a result of high-angle faulting that created discrete peridotite ridges. In this case the recovery at Site 1277 could represent mass wasting of material from such a serpentinite ridge into a local low (i.e., small basin) that was elevated well above the regional seafloor and therefore isolated from terrigenous sources. Mass wasting of an upfaulted serpentinite ridge might have resulted in a vertical sedimentary organization of the detritus (e.g., a fining-upward sequence as the ridge was denuded); however, this is not observed and, in fact, the largest known detached block (meter sized) is located toward the top of the recovery. Site 1277 serpentinite breccias also clearly originated as multiple debris flows rather than as massive slides from serpentinite fault scarps, as inferred for some of the Iberia breccias (Gibson et al., 1996).
The first (i.e., continuous) process is consistent with Alpine counterparts because serpentinite breccias seem to be a near-ubiquitous feature wherever exhumed mantle has been identified, rather than merely forming along discrete serpentinite ridges. The Alpine serpentinite breccias appear to have formed a laterally persistent blanket that was genetically linked with mantle exhumation along a gently inclined (<20°) master detachment. The most probable explanation for the absence of terrigenous sediment at Site 1277 is that the upper plate of detached continental crust was already far removed (several kilometers or more) before the serpentinite mass flows were deposited. The area of active mantle exhumation was possibly isolated from terrigenous input by topographic barriers (e.g., other mantle fault blocks) between Site 1277 and the rifted margin to the west. In continental core complexes (e.g., Aegean Turkey) detrital sediment overlying inferred low-angle extensional faults is similarly derived from the exhumed lower plate (i.e., high-grade metamorphics), with little or no material from the detached upper plate (i.e., lower grade or unmetamorphosed rocks) (e.g., Purvis and Robertson, 2004).
Two main alternatives for the origin of the basaltic volcanism at Site 1277 are considered. The first is that the MORB from the Newfoundland margin formed in a setting of very slow seafloor spreading. However, the lava flows interbedded with serpentinite mass flows show that this basalt does not represent detritus from oceanic crust that was otherwise not recovered. The favored second alternative is that the MORB erupted following the completion of mantle exhumation and reflects the earliest basaltic volcanism related to the beginning of seafloor spreading. Assuming this is correct, sampling at Site 1277 was fortuitous; a drill location that reached basement farther west (continentward) would have sample only exhumed mantle lithosphere or continental crust, whereas a location farther east (oceanward) would have sampled "normal" oceanic crust.
The relatively depleted melt represented by the MORB volcanics is compatible with the extraction of melt from more fertile mantle than the depleted mantle harzburgite of the basement at Site 1277 (Müntener and Manatschal, 2006). One possible explanation is extreme mantle heterogeneity (i.e., highly depleted vs. undepleted upper mantle sources in a small area). Another possibility is that different slices of crust and mantle existed at different depths within a zone of rift-related extension along the Newfoundland margin. This possibly included ophiolitic mantle (i.e., the depleted ultramafic basement), continental crust (influencing the Th and Hf ICP-MS analysis), and deep "normal" upper mantle that is the source of the MORB. The Newfoundland margin is made of several tectonic terranes that were accreted at different times, notably the Avalon Terrane, which was emplaced during the Silurian–Devonian, and the Meguma Terrane, which was emplaced during the Carboniferous–Permian (Jansa and Wade, 1975; Williams et al., 1995). In addition, Hercynian-age (Carboniferous–Permian) lithosphere may be present in the vicinity of the Newfoundland margin, as suggested by the ages of detrital muscovites recovered at nearby Site 1276 (Wilson and Hiscott, this volume). In general, the presence of continental crust, mantle of different compositions, and small differences in the degree of partial melting could all have contributed to the observed chemical patterns of the Site 1277 basalts.
The brittle deformation represented by subvertical fissures and veins in both the peridotite basement (Unit 2) and the mass flows (Unit 1) clearly postdated the ductile high-temperature deformation and exhumation. Several alternative scenarios are considered.
One possibility is that the brittle deformation represents the late stage of a single phase of extensional deformation (e.g., related to bending stresses that developed in the footwall during exhumation). Early high-temperature ductile deformation could have given way to brittle deformation during such progressive exhumation and cooling.
A second possibility is that the brittle fracturing reflects a separate phase of extension that followed initial mantle exhumation. This is favored for several reasons. First, faulting and veining in the serpentinite breccias could only have been preserved as brittle deformation structures after at least partial lithification of the serpentinite breccias and interbedded finer-grained sediments, which suggests the existence of a time gap between basement exhumation and high-angle brittle deformation. Second, significant tilting would be expected during the brittle phase of progressive exhumation. However, the subtle primary lamination in the fine-grained sediments filling the fissures (i.e., geopetal fills) shows that relatively little tilting about a vertical axis (<20°) has taken place since the fissures opened and were filled with sediment. Third, evidence from the Iberia margin (Comas et al., 1996) and from the Alps (Manatschal et al., in press) suggests that a discrete phase of brittle deformation did indeed follow mantle exhumation in those areas. Such late-stage high-angle brittle deformation might reflect a continuation of the same tensional regime that controlled the exhumation. For example, in continental core complexes, early low-angle faults have been abandoned and replaced (i.e., cut through) by later high-angle faults when it became more energy efficient to generate new high-angle faults (Profett, 1986). On the other hand, the high-angle brittle structures might relate to a later, unrelated phase of crustal extension (see below).
A third possibility is that the high-angle structures might reflect gravity sliding or mass flow at the seafloor following exhumation. This interpretation would require the sliding and related internal deformation of a coherent mass of debris. However, gravity-driven structures in such debris might be variably oriented, listric, or low angle in geometry, in contrast to the consistent subvertical orientation observed at Site 1277.
The interpretation favored here is that the brittle fracturing and extension reflect the latest phase of extension related to mantle exhumation. This took place after exhumation of mantle along low-angle faults but possibly before the onset of "normal" seafloor spreading.
Several modeling studies have suggested that the initiation of "normal" mid-ocean-ridge spreading coincides with the release of in-plane tensile stress (i.e, relative compression) as large volumes of magma enter the rift zone (e.g., see Tucholke et al., 2007, for discussion). The modeling implies a long-wavelength lithosphere flexural response, which could be difficult or impossible to recognize using drilling data. Localized compression might also occur along weak zones in basement, resulting in local uplift, reverse faulting, or inversion of preexisting extensional faults. From this point of view, the Unit 2/1 boundary at Site 1277 represents an obvious weak zone located close to the locus of final breakup of subcontinental lithosphere, and it might possibly record a "relative compression event." The sediment infills of the high-angle fissures in fact record incremental widening related to progressive extension. Similar late-stage brittle extension was recorded on the Iberia margin (Beslier et al., 1996; Cornen at al., 1996; Morgan and Milliken, 1996). The proposed compression (late Aptian) postdates the normal faulting by as long as 14 m.y. and could therefore have postdated the end of extension documented by the veins (see Tucholke et al., in press). However, some the veins remain unfilled as open fractures and represent persistent zones of weakness. These open fractures could have collapsed if regional compression was locally expressed at Site 1277.
Site 1277 was drilled close to the crest of the Mauzy Ridge (Fig. F2). It is uncertain when and how the ridge was uplifted. Deep seismic reflections lap against the Mauzy Ridge (see Fig. F2 in Shipboard Scientific Party, 2004a), showing that the ridge formed during or not long after the final stages of continental breakup and that it is not likely to represent a geologically young feature. Two main possibilities can be considered.
One alternative is that the serpentinite ridge formed within the ocean–continent transition zone during exhumation of the mantle. Bending stresses associated with large-scale detachment faulting are thought to have broken the footwall and uplifted seafloor ridges to hundreds of meters high on the Mid-Atlantic Ridge (Tucholke et al., 1998). As noted earlier, continental core complexes are also known to undergo high-angle faulting following initial exhumation in areas such as the western United States (Profett, 1988) and western Turkey (Purvis and Robertson, 2004). This faulting is integral to progressive extension.
A second alternative is that the Mauzy Ridge was uplifted by normal faulting well after the mantle was first exhumed (i.e., during a later phase of extension), as proposed by Peron-Pindivic et al. (2007). The MORB lavas are on the crest of a long narrow ridge, elevated several hundred meters above the surrounding seafloor. This suggests that the ridge could have formed after clastic sedimentation and volcanism ended at Site 1277. The lavas were erupted on only gently inclined seafloor; eruption on a slope would have created pillow breccias or pillow lavas rather than the massive flows that dominate the core recovery. Extension is expected to have ended when the subcontinental mantle lithosphere finally separated and large volumes of magma entered the rift, initiating "normal" seafloor spreading. It has been proposed that this may have occurred as late as the late Aptian to early Albian (Tucholke et al., 2007). Thus, available evidence indicates the following progression of events:
The time at which each of these events occurred and the question of their separation or overlap remains to be resolved by further drilling.