SCIENTIFIC RATIONALE FOR DRILLING

Atlantis Massif has several key features that make it an ideal target for OCC drilling. The massif is <2 Ma, so weathering and erosion have not degraded (macro)structural relationships. The hanging wall is in contact with the footwall (core) of the detachment fault. The core of the massif is dominated by variably serpentinized peridotite at the surface, and mantle seismic velocity has been reported to occur at several hundred meters depth below seafloor, affording access to fresh in situ mantle with conventional drilling; therefore, we can document the alteration gradient and front with depth. Our scientific objectives cannot be accomplished via seafloor mapping/sampling.

The hypotheses that drilling results will test are listed below, followed by the specific observations/measurements expected to be most critical. Whereas many of the observations and measurements outlined below will be initiated at sea, most of the results will be ascertained only after postcruise research programs.

• thickness of the fault zone
• ductile-brittle evolution
• temperature and pressure conditions of equilibrium mineral assemblages
• relationships between deformation and recrystallization
• history of strain localization within the fault zone

If core is recovered from within the detachment fault zone we will use micro- and macrostructural measurements to estimate the sense and magnitude of slip that has occurred on single slip surfaces and detail the occurrences of multiple slip planes that may work together to accommodate extension/plate spreading. The volume, structure, and composition of alteration products within the fault zone will provide constraints on slip accommodation and fluid flow along the fault. Observations of the lower boundary of the tectonized zone will be important for understanding whether strain is isolated or more distributed throughout the rocks some distance below the detachment (e.g., Fletcher and Bartley, 1994). Logging data will play a key role in understanding the detachment fault zone because recovery is likely to be poor due to the highly fractured nature of the rock. Downhole changes in fracture patterns, porosity, resistivity, and magnetic susceptibility will aid discrimination of fault zone structure and its variation with depth.

The history of exposure of the detachment fault at the seafloor may be reflected by paleoceanographic indicators within sediments that drape the footwall. Sedimentation rates at 30°N in the Atlantic Ocean are sufficiently high that basement rock exposure on the dome is quite rare. Recovery of the thin (~1 m) lithified sediment section may provide constraints on the timing of footwall exhumation, which occurred over an unknown period in the past 1–2 m.y. In a given location, recovery of the topmost deposits would indicate microfossil and oceanographic (temperature and stable isotope) signatures characteristic of the deposition time. Recovery from a series of relatively shallow or bit-to-destruction spreading-parallel sites spanning the central dome could provide limits on the rate of unroofing (systematic changes in indicators from west to east) and perhaps on the time of initial fault exposure (age of the sediments on the western edge of the dome). Comparison of the top sediments at the central dome to those on the southeast shoulder could address questions about differing evolution of these two portions of the detachment fault system (Blackman et al., 2004).

• depth distribution of deformation degree/style
• localization of deformation
• alteration concentrated in local zones
• brittle deformation dominant in hanging wall
• rotation of tectonic blocks

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). In this scenario, the pattern of footwall strain should vary from extension near the upper surface, through an interval of no evident strain, to compression in the lower part of the plate. Differential rotation between the footwall and hanging wall blocks is predicted by thin-plate theory, so we will investigate whether the core or logging data show evidence of such history. Logging data will provide continuous (orientated) images of fracture patterns in the borehole wall. These data will be compared with fractures and veins measured in the cores from the same depth interval. With constraints on core orientation and logging tool calibration, paleomagnetic data can be related (MacLeod et al., 1994, 1995) to any systematic rotation of the footwall and hanging wall.

• fluid inclusion analyses
• metamorphic history
• thermochronometry and thermal evolution

Gabbro is interlayered with peridotites along the south wall of the massif, and the core of the massif is likely composed of similar rock types. Fluid inclusions within gabbro can be used to determine the depth at which the rock formed (i.e., the pressure and temperature at which the fluid was entrapped) (e.g., Kelley and Delaney, 1987; Kelley and Früh-Green, 2001). Comparison of such data with the present depth of core samples can provide limits on the amount of uplift/unroofing. The pressure/temperature evolution of metamorphism will reflect the tectonic evolution as well, with cooling rates and water/rock ratios being controlled by the amount of unroofing together with degree of fracturing. The detachment model predicts that the hanging wall basalts initially overlay the footwall. If this is the case, petrological and geochemical results are expected to show a genetic relationship between footwall peridotites, any melt residue therein, and the basalts of the hanging wall. Rock recovered via coring will allow this prediction to be tested.

• distribution of alteration above serpentinization front
• orientation of deformation fabrics at angles to plate spreading direction
• evidence for significant rock-seawater exchange
• timing of alteration with respect to deformation
• microstructural analysis

If expansion of serpentinized peridotite contributed significantly to the evolution of the massif, the bulk density of the altered material must be low enough to allow it to move relative to the surrounding rock. Plastic flow structures within the serpentinite might reflect significant vertical shear, due to buoyancy of the expanding material.

• link seismic velocity gradient to vertical distribution of rock types
• degree/distribution of alteration and cracking

Serpentinization of olivine in lower crustal and upper mantle rock types is associated with the uptake and release of both major and minor elements and compounds, including H₂O, Mg, Ca, Si, Cl, and B, which has important consequences for long-term global geochemical fluxes. Besides the production of heat through exothermic reactions, serpentinization leads to reduced, high-pH fluids with high H₂ and CH₄ concentrations. Recovery of this transition will allow quantitative modeling of geochemical changes associated with progressive serpentinization. As yet, the end-member compositions, and therefore geochemical fluxes associated with serpentinization, have been necessarily inferred from primary mineral modes and compositions.

• variations in degree of melting, distribution of retained melt, and/or melt-rock reaction as functions of depth
• final depth of melting
• relationship between magmatism and subsolidus deformation

The processes responsible for the development of oceanic core complexes appear to be episodic, with one factor being the level or style of magmatic activity at the local spreading center. Evaluation of the degree of mantle melting the peridotites have undergone can be estimated from reconstructed residual mode and bulk rock (major element) chemistry and from major and trace element mineral chemistry (Cr# in spinel and pyroxenes, Mg# in olivine, TiO₂ in orthopyroxene and clinopyroxene, and heavy rare earth elements in clinopyroxene [Hellebrand et al., 2001]). Melts retained within residual peridotite may reflect in situ partial melting (Elthon, 1992) or transient melts ascended from deeper levels. This melt may react extensively with or otherwise refertilize the peridotite, altering both melt and residue compositions (e.g., Daines and Kohlstedt, 1994; Edwards and Malpas, 1996; Elthon, 1992; Johnson and Dick, 1992; Kelemen et al., 1992, 1997; Seyler and Bonatti, 1997; Seyler et al., 2001). Dissolution textures and mineral inclusions allow evaluation of the nature and amount of melt-rock interaction.

• core lithology and degree of alteration
• presence/lack of significant iron oxide gabbro
• correlation with seismic velocity

Late-stage magmas containing significant proportions of Fe and Ti can be emplaced in the crust at irregular intervals (Agar and Lloyd, 1997; Natland et al., 1991), thereby increasing the overall density of the intruded body. In this case, both high and average (crustal) density material might be mainly gabbroic, but they would have different bulk properties. To date, seafloor samples from Atlantis Massif do not indicate that high densities evident in the MAR 30°N residual gravity (Fig. F3) (Blackman et al., 2004) are due to the presence of significant Fe-Ti gabbro. Although the densities of such oxidized rocks can be high, the seismic velocity is not increased correspondingly (Itturino et al., 1991; Miller and Christensen, 1997), so Fe-Ti gabbro would not cause the high velocities observed. However, only through drilling will it be possible to determine what rock types and their relative degree of alteration generate the observed geophysical signals.

Qualitative inferences derived from seafloor mapping and sampling at Atlantis Massif can become quantitative constraints on models when drill core and logging measurements/analyses are available. Key insights into uppermost mantle processes and the effect of alteration on rheology and geochemistry will be obtained through comparison of core samples from this site and those drilled at ODP Sites 670, 920, and 895 and sites of Leg 209. If recovery at the footwall site is sufficient, we will obtain information on depth variation in preferred alignment of minerals associated with flow in the mantle; together with the distributed, shallow samples obtained during Leg 209, analysis of any systematics of fabric with respect to tectonic setting can be addressed. Data from Expeditions 304 and 305 may also contribute to such studies directly if contingency operations obtain core from site(s) offset along strike from the priority footwall site.

Data from Expeditions 304 and 305 will also be used to address what controls fault geometry and slip behavior at different types of OCCs. Drilling results from Atlantis Bank (Hole 735B) (Dick et al., 2000) and seafloor mapping/sampling results from the 15°45'N MAR complex (MacLeod et al., 2002; Escartin et al., 2003) suggest that the depth dependence and relative timing of brittle versus ductile deformation may differ between OCCs. New information from Godzilla Megamullion (Ohara et al., 2001, 2003) is also expected to be directly relevant to this question. Drilling results from Atlantis Massif will provide key data for comparison.