BACKGROUND AND OBJECTIVES

Atlantis Massif formed within the past 1.5–2 m.y., and it currently bounds the median valley on the western flank of the at Mid-Atlantic Ridge (MAR) 30°N (Fig. F1). The corrugated, striated central portion of this domal massif displays morphologic and geophysical characteristics inferred to be representative of an oceanic core complex (OCC) exposed via long-lived detachment faulting (Cann et al., 1997; Blackman et al., 1998, 2004; Collins et al., 2001). The "core" of the complex is composed of crust and possibly upper mantle rocks, denuded by a detachment fault exposed over an 8–10 km wide, 15 km long area that forms the elongate, doubly plunging domal seafloor morphology. An adjacent basaltic block to the east is interpreted as the hanging wall to the detachment fault. A thin cover of lithified sediment, volcanic deposits, and rubble on the dome of the massif impedes seafloor mapping and sampling of the fault surface. The sediment-draped volcanic morphology and basalt sampled from scarps in the eastern block show its general character; the location of its contact with the dome can only be inferred from the break in slope. The detachment is inferred to dip beneath the seafloor at the base of the dome and to continue at a shallow angle (<15°) beneath the eastern block.

Evolution of the southern portion of the massif may differ from that of the central dome. The southern ridge (Fig. F1) has experienced greater uplift, shoaling to 700 m below sea level (mbsl). There, the corrugated surface extends eastward to the top of the median valley wall. Exposures along the south face of the massif represent a cross section through the core complex. The serpentinization-driven Lost City hydrothermal vent field is just below the summit of the southern ridge (Kelley et al., 2001, 2003; Früh-Green et al., 2003).

Seismic refraction results at Atlantis Massif (Fig. F2A, F2B) (Detrick and Collins, 1998) indicate that velocities of 8 km/s occur within several hundred meters of the seafloor at least locally in parts of the core of the massif. The gradient of seismic velocity in the central dome of Atlantis Massif has been determined to be similar to that determined near Ocean Drilling Program (ODP) Site 920, where 100–200 m of serpentinized peridotite was drilled. The determined gradient is quite distinct from that characterizing gabbro-hosted Atlantis Bank (Southwest Indian Ridge) and other sections of the MAR.

Interpretation of multichannel seismic (MCS) reflection data suggests a major difference in structure between the outside (conjugate) corner lithosphere versus that hosting Atlantis Massif (Canales et al., 2004). The Layer 2a/2b boundary is quite clear on the eastern flank of the ridge axis, but it is not evident on the western flank across the massif. A strong reflector is visible at 0.2–0.5 s below much of the domal surface (Fig. F2C, F2D) and coincides roughly with the depth below which mantle velocities (~8 km/s) are deduced. One interpretation suggests that the reflector marks an alteration front within the peridotite-dominated massif (Canales et al., 2004).

Modeling of sea surface and sparse seafloor gravity data (Blackman et al., 1998, 2004; Nooner et al., 2003) suggests that rocks at depth beneath the dome have a density 200–400 kg/m³ greater than the surrounding rock. Two-dimensional model results support the interpretation that the footwall is overlain by tilted hanging wall blocks capped by rocks with density typical of the upper crust (2.5–2.7 kg/m³). The interface between the model blocks in the east is a gently inclined (15°–25°) boundary that dips more steeply than the exposed corrugated surface (~11°) where it meets that hanging wall, perhaps coinciding with the base of the fractured, highly altered detachment fault zone.

In situ rock samples from scarps, side-scan imagery, and gravity data suggest that the majority of the hanging wall block is composed of erupted basalt. Seismic data show a discontinuous but persistent reflector 0.2–0.5 s beneath the seafloor, which Canales et al. (2004) show coincides with the projection of the corrugated slope beneath the western edge of the hanging wall block. They interpret this reflector to be the unexposed detachment fault. Assuming an average velocity of 4 km/s in fractured basalt, the reflector is predicted to occur at 200–300 m below seafloor (mbsf) at the hanging wall drill site.

Rock samples collected by the manned submersible Alvin from the central dome are mostly angular talus and rubble of serpentinized peridotite, metabasalt, and limestone (Cann et al., 2001; Blackman et al., 2004). A few samples from the central dome show cataclastic deformation or are highly serpentinized and metasomatically altered peridotite. The protolith of most of the serpentinite sampled on the south wall is harzburgite. These rocks are commonly cut by highly altered gabbroic veins composed dominantly of talc, tremolite, and chlorite (Früh-Green et al., 2001; Schroeder et al., 2001). Low-temperature overprinting, seafloor weathering, and carbonate vein formation mark the youngest phases of alteration.

Microstructural analysis of samples from the south wall indicates shear deformation and dilational fracturing at metamorphic conditions ranging from granulite to sub-greenschist facies (Schroeder et al., 2001). Ductile fabrics in peridotite samples are overprinted by semibrittle and brittle deformation (Schroeder and John, 2004). Stable mineral assemblages of tremolite, chlorite, and chrysotile indicate that the latter processes occurred at temperatures <400°C. The distribution of samples suggests that strong semibrittle and brittle deformation is concentrated at shallow structural levels (<90 m beneath the domal surface) at the southern ridge (Schroeder and John, 2004). Outcrop mapping with the Alvin and photomosaics constructed from Argo digital still camera shots show that this uppermost fault extends across much of the top of the southern ridge (Karson, 2003).

Atlantis Massif has several key features that make it an ideal target for OCC drilling: it is less than 2 m.y. old, so weathering and erosion have not degraded (macro) structural relationships; the hanging wall is in contact with the footwall of the detachment; and mantle seismic velocity has been reported to occur at several hundred meters below seafloor of the domal core, affording access to fresh in situ peridotite with conventional drilling. Many fundamental questions about mantle melting can only been addressed in a limited way with the highly altered samples of oceanic peridotite available to date. If the geophysical data indicate the presence of residual mantle and drilling can recover these rocks, new insights would be gained on geochemical balances during melting, the nature of melt segregation and migration, and the rheology of the asthenosphere within the subaxial upwelling zone. This was a very high priority for the combined Expeditions 304 and 305, and the operational strategy was designed to maximize the chances for successful recovery of deep footwall rocks.

Hypotheses to be tested by drilling completed during Expeditions 304 and 305 are:

Expedition 304 results allow testing of the first 4 hypotheses and Expedition 305 results are expected to address these as well as the last 3 hypotheses.

If long-lived normal faulting and displacement are responsible for the evolution of the massif, uplift of the core may be the result of isostatic adjustment (Vening Meinesz, 1950) and thin-plate flexure (Spencer, 1985; Wernicke and Axen, 1988; Buck, 1988; Lavier et al., 1999). Differential rotation between the footwall and hanging wall blocks is predicted by thin-plate theory, so we apply the Integrated Ocean Drilling Program (IODP) results to investigate whether the core/logging data show evidence of such history. Logging data provide continuous (oriented) images of fracture patterns in the borehole wall. These are compared with fractures and veins measured in the cores from the same depth interval. Paleomagnetic data are incorporated to determine any history of rotation of the upper footwall. The pressure/temperature evolution of alteration reflect the tectonic and magmatic history as well, with cooling rates and water/rock ratios being controlled by intrusions, the amount of unroofing, and the degree of fracturing.

Any detachment model predicts that hanging wall rocks initially reside structurally above the footwall. If this is the case, petrologic and geochemical results are expected to show a genetic relationship between footwall rocks and basalts of the hanging wall.

The processes responsible for the development of OCCs appear to be episodic, with one factor being the level or style of magmatic activity at the local spreading center. Detailed study of the igneous sequence and structural relationships therein will be used to address the evolution of melting, intrusion, and cooling during the formation of Atlantis Massif. Comparison of our findings with those from ODP Hole 735B, Leg 153, and Leg 209 provide a means for assessing the similarities and differences in conditions that prevailed at the (slow) spreading centers where the lithosphere drilled at these sites was initially formed.