SUMMARY

The footwall exposed on the domal core of Atlantis Massif is heterogeneous. Site U1309 is located on the southern part of the central dome. Peridotite and the Lost City vent field (interpreted to be serpentinization-driven) crop out south of Site U1309 along the southern ridge of the massif (Kelley et al., 2001). Exposure of this apparently significant volume of peridotite 5 km south of Hole U1309D contrasts strongly with the ~1.4 km of gabbroic rock recovered at Site U1309. Analysis of the NOBEL seismic refraction data, centered ~2 km north of Hole U1309D, indicates that rock with a seismic velocity > 7.5 km/s, presumably dominantly olivine and essentially unaltered, is present at ~800 mbsf. This is in contrast with the ubiquitously lower velocities that typify gabbroic rocks such as those that dominate the recovery from Hole U1309D.

Initial work relating drilling results to possible broader scale geophysical signatures suggests some correspondence between rock type and alteration and seismic properties, but structural controls, not surprisingly, also appear to play a role. The ~1.4 km sequence of dominantly gabbroic rocks interlayered with 1–25 m thick ultramafic rocks or dunitic troctolite recovered from Hole U1309D clearly indicates that Atlantis Massif is not simply an uplifted mantle section where serpentinization is responsible for lower densities/seismic velocities (only) in the upper few hundred meters. A more complex model than the one put forward before Expeditions 304 and 305 (Cann et al., 1997; Blackman et al., 1998; Blackman, John, Ildefonse, MacLeod, Ohara, Miller, and the Expedition 304/305 Project Team, 2004; Collins et al., 2001; Canales et al., 2004) is required. The lack of recovery of fresh mantle peridotite, or even significant sections of serpentinized mantle peridotite, is in contrast with exposures of serpentinized mantles along the southern ridge of the massif. This requires any models we develop to incorporate complex lateral and vertical heterogeneity in lithography, alteration, and structures.

Onboard synthetic seismogram modeling (Fig. F35) suggests that physical property changes across the 300–400 mbsf interval could give rise to the strong D reflector (Canales et al., 2004) (Fig. F3). 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 in the upper few hundred meters of the massif. Given the lithologic (and related alteration) variability with depth in Hole U1309D, it would be interesting if the interface between gabbroic rocks and dunitic troctolite actually extends at a similar depth across the central dome of Atlantis Massif. Although still associated with a contrast in alteration, the increase in alteration with depth and the fact that this interval is embedded within a much larger gabbroic sequence contrast sharply with the interpretation (Canales et al., 2004) of the D reflector as the base of a regional alteration front.

The gabbroic rocks sampled from Site U1309 are among the most primitive known along the entire MAR. The most olivine-rich end-member consists in a series of dunitic troctolite intervals, on average moderately serpentinized and locally very fresh. The preferred shipboard interpretation is that they constitute the primitive cumulate of the recovered igneous sequence(s). Oxide gabbros, a common feature in slow-spreading ridge boreholes (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. Commonly, mylonitic shear zones overprint these oxide-bearing intervals. However, they are most common 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. The most significant unit boundary within the recovered gabbroic pluton seems to be located at ~600–800 mbsf, as indicated by a series of observations, including a rapid change in the geochemical compositions at ~600 mbsf, a series of faults with gouges between ~ 695 and 785 mbsf, and a change in the late metamorphic overprint below ~800 mbsf.

There is a striking lack of extensive amphibolite facies alteration and deformation in rocks from Site U1309 in the same way as is lacking at the 15°45'N corrugated dome on the MAR (Escartin et al., 2003). This contrasts markedly with the nature of the gabbroic section recovered from ODP Hole 735B at the Southwest Indian Ridge (Robinson, Von Herzen, et al., 1989; Dick, Natland, Miller, et al., 1999; Dick et al., 2000). Overall, the rocks recovered from Hole U1309D show little deformation, suggesting that large-scale deformation associated with a detachment fault was either not recovered in the upper section of the borehole at Site U1309 or occurred at low temperature. The lack of extensive deformation in these footwall rocks suggests that any deformation related to a major detachment fault system must have occurred at low temperature and must be strongly localized. IODP drilling results at Atlantis Massif are consistent with the low-temperature shallow rooting detachment fault model proposed by MacLeod et al. (2002) to have controlled the evolution of the corrugated dome west of the MAR at 15°45'N.

Hole U1309D is located approximately midway downdip along the exposed corrugated detachment fault surface. The breakaway to this system is inferred to be ~5 km west. The fault termination (where the fault dips below the basaltic hanging wall block) lies ~5 km east. The dip of the fault at its termination is ~11°, and, therefore, the minimum amount of rotation expected for this site is of the same magnitude, larger if the detachment originally steepened toward the ridge axis. The shipboard paleomagnetic measurements indicate that no significant net rotation ( 15°) of the footwall rocks with respect to the predicted geomagnetic field has occurred at this site below the Curie temperature (~520°–580°C). This lack of footwall rotation suggests that a rolling hinge model is not a viable explanation for the uplift of the core of Atlantis Massif along a single concave, normal fault. A model involving multiple faults is probably more likely correct. The low-angle detachment fault currently capping Atlantis Massif could have captured the recovered gabbroic section at a relatively shallow depth in the lithosphere. As proposed previously (e.g., Karson, 1990; Cannat et al., 1997; Lagabrielle et al., 1998; Kelemen, Kikawa, Miller, et al., 2004), the pluton may have first risen from its deeper, crystallizing depth along a series of conjugate normal faults.