Previous   |    Next

doi:10.2204/iodp.proc.304305.101.2006

Discussion

Site U1309 is located at the surface of a low-angle detachment fault exposed at the seafloor on Atlantis Massif. Holes U1309B and U1309D are at the base of a north-facing, ~12° slope running down from the Southern Ridge to the central dome, and the drilling took place close to the point of inflection (in the spreading-parallel direction) of the domal fault surface. The breakaway of the fault is inferred to be ~5 km to the west, and the fault termination (where the exposed fault intersects the seafloor) ~5 km to the east. The dip angle of the fault at its termination is ~11°; this value provides a minimum estimate for footwall rotation at Site U1309 because the fault is approximately horizontal at this location. Shipboard paleomagnetic data show that from 0 to 180 mbsf in Hole U1309D the remanent vector inclination is not distinguishable from the expected dipole inclination at this site. Mean inclinations from the paleomagnetically defined groups (II–V) deeper than 180 mbsf differ from the local geocentric dipole inclination and require at least moderate and variable rotation between the different blocks. Interpretations of the amount of tectonic rotation of the recovered sequence are speculative at this stage, but it is clear that steady rotation of a single, ductilely deforming footwall block cannot explain the structural and paleomagnetic data.

The gabbroic section recovered from Site U1309 shows a lack of extensive amphibolite-facies alteration and deformation. A similar relationship is noted in shallow-penetration cores collected from the corrugated dome at 15°45′N on the MAR (MacLeod et al., 2002; Escartin et al., 2003) and contrasts markedly with the gabbroic section recovered from Hole 735B at the Southwest Indian Ridge (Robinson, Von Herzen, et al., 1989; Dick, Natland, Miller, et al., 1999; Dick et al., 2000). The overall low intensity of deformation at Site U1309 suggests that structures associated with significant slip on the detachment fault were either localized within the unrecovered upper 20 m or occurred at low temperature within the brittle regime and are distributed along faults documented within the upper ~250 m of the footwall.

The relatively little-deformed nature of the plutonic section recovered from Site U1309 allows an unprecedented opportunity to study emplacement processes associated with formation of oceanic lithosphere. Intrusive contacts are preserved in many places, and their dip may suggest the type of injection (sill versus dike, for example). The observations of gabbro dikes and late magmatic leucocratic veins that crosscut more primitive rock types indicate that melt migration was controlled by brittle mechanisms during the later stages of the fractionation and crystallization process. The observation of crystal-plastic shear zones within narrow intrusions/dikes and at the contacts between gabbroic intervals suggests that the presence of melt promoted strain localization at/near the ridge axis.

The gabbroic rocks sampled from Site U1309 are among the most primitive known along the entire MAR. The most olivine-rich end-member consists of a series of moderately serpentinized, olivine-rich troctolite intervals, locally fresh, that may represent primitive cumulates. Oxide gabbros, a rock type commonly recovered from slow-spreading mid-ocean ridges (Robinson, Von Herzen, et al., 1989; Dick, Natland, Miller, et al., 1999; Pettigrew, Casey, Miller, et al., 1999; Kelemen, Kikawa, Miller, et al., 2004) are also present at Site U1309. Mylonitic shear zones can overprint these oxide-bearing intervals. However, the latter are most common in Hole U1309D in undeformed rocks with magmatic textures and either sharp or diffuse boundaries. The interplay between relatively late Fe-Ti oxide crystallization and deformation is probably complex, and their relative timing may be variable.

There is little magmatic deformation associated with the intrusive igneous history. This observation suggests that plutons were possibly “shielded” from deformation related to emplacement within the mantle lithosphere. Alternatively, the lack of significant magmatic deformation may simply reflect differences in the timing of deformation and plutonism at the ridge axis. The drilled gabbro section is relatively continuous and structurally homogeneous, except for zones of increased crystal-plastic and brittle deformation in the upper ~320 mbsf and between 650 and 800 mbsf. Following granulite-grade shearing, there is relatively little evidence for deformation at amphibolite facies. Evidence for deformation related to denudation in the shallow part of Holes U1309B and U1309D suggests that faulting initiated at greenschist-grade conditions. These observations suggest that the footwall cooled considerably between pluton emplacement and denudation at the seafloor.

The most significant boundary within the recovered gabbroic sequence appears to be located at ~600–800 mbsf, as indicated by a series of observations, including a marked change in Mg# at ~600 mbsf (Fig. F15), a series of faults with cataclasite between ~685 and 785 mbsf (Fig. F29), changes in paleomagnetic inclination, and a change in the late metamorphic overprint below ~800 mbsf (Figs. F17, F34).

The regional seismic reflector D was interpreted by Canales et al. (2004) to correspond to an alteration front in mantle peridotite or to a deeper detachment fault within a system of normal faults that controlled the evolution of the OCC. At the drill site, D-reflector could correspond to a strong reflector at ~0.12 s (two-way travel time [TWT]) or to a reflector near 0.25 s TWT (see Figs. F2, F25). However, neither of these reflectors corresponds to an alteration front in mantle peridotite or a clearly defined deeper detachment fault in the recovered core. Complexity along the strike of MCS Line Meg-4 makes it difficult to be certain; the stacked section (not shown) suggests that 0.12 s marks D-reflector, but the migrated section (Fig. F3) suggests that it occurs at 0.25 s near Hole U1309D.

Onboard synthetic seismogram modeling (Fig. F25) indicates that physical property changes across the 300–400 mbsf interval could give rise to D-reflector. If ~340 mbsf coincides with D-reflector (~0.12 s TWT), the average velocity of the overlying section would be ~5.44 km/s locally, slightly lower than the 5.54 km/s value determined by a check shot to 345 mbsf. Both of these values are higher than the 5.0 km/s noted by Canales et al. (2004) for the interval velocity above this reflector in the center of MCS Line Meg-10. In Hole U1309D, it is probably the change from less altered gabbro to more altered olivine-rich rocks that dominates the local acoustic impedance contrast. Given the lithologic (and related alteration) variability with depth in Hole U1309D, it would be surprising if the ~300–400 mbsf interfaces between gabbroic rocks and olivine-rich troctolite extend at a similar depth across the central dome of Atlantis Massif. The increase in alteration in the interval 280–340 mbsf and the fact that it is embedded within a much larger gabbroic sequence contrast sharply with the interpretation (Canales et al., 2004) of D-reflector as the base of a regional alteration front.

If, on the other hand, D-reflector in Hole U1309D coincides with the ~0.25 s TWT reflector, the changes in physical properties (drop in porosity, jump in resistivity, and drop in alteration; see “Geophysical measurements”) across the 700–800 mbsf interval might explain the impedance contrast on a seismic wavelength. However, we do not have sonic log data to confirm in situ velocities below 750 mbsf. Tying D-reflector to this deeper set of physical property/​alteration changes could support Canales et al.’s (2004) inference of a subsurface fault giving rise to D-reflector. Fault gouge recovered near 750 mbsf confirms the existence of at least a narrow fault zone; low recovery over a broader interval (10–20 m) may indicate a greater width. The dip in borehole temperature at this depth suggests the fault zone could be permeable enough to allow fluid flow, and such localized change in properties would certainly contribute to an observable impedance contrast.

Our findings at Site U1309 are in marked contrast with the expectation that we would drill through an alteration front in mantle peridotite at Atlantis Massif. The results are inconsistent with the predrilling hypothesis that the footwall was composed mainly of an uplifted mantle section where serpentinization was responsible for lower densities and high seismic velocity gradients in the upper few hundred meters of the footwall. A more complex model than was put forward before Expeditions 304 and 305 will be required. The fact that we did not reach fresh mantle peridotite, together with the known exposures of serpentinized mantle along the southern face of the massif, supports models of significant lateral heterogeneity in slow-spreading oceanic crust. The drilling results indicate that more detailed analysis of the existing geophysical data is warranted, using more intensive processing methods and testing a suite of possible three-dimensional subsurface models. Together with the core and logging data from Hole U1309D, this work will provide an unprecedented opportunity to advance understanding of lower crustal accretion and tectonic evolution at slow-spreading ridges.

Top of page   |    Previous   |    Next