Present models for the ocean crust also suggest that there are radical differences in the stratigraphy of the lower crust according to spreading rate, proximity to hot spots, local ridge geometry, and proximity to fracture zones. Demonstrating the extent to which these differences are genuine and how they are manifested in the lower ocean crust are major objectives of the proposed lower crustal and shallow-mantle drilling program of the internationally coordinated effort to study the accretion of the ocean crust at mid-ocean ridges (InterRidge). This involves both drilling a representative section of slow-spread lower crust at Site 735, and further offset drilling in tectonically exposed East Pacific Rise lower crust to obtain a composite section for the fast-spread end-member.
Deep crustal sections exposed in ophiolites are highly variable, both between different ophiolites (e.g., Coleman, 1977) and within a single ophiolite (e.g., Nicolas et al., 1996). Because many ophiolites are associated with the formation of ocean crust in island-arcs, fore-arcs, back-arc basins and small oceanic rift basins, it also is not known to what extent they provide a valid analog to modern ocean crust formed beneath the major ocean basins. The criteria for discriminating primary tectonic variables, such as spreading rate and proximity to transforms, are also largely conjectural. This section of ultra-slow-spread lower crust provides objective criteria for evaluating the provenance of individual ophiolite complexes because its proximity to transform offsets and its position in the paleo-ridge segment are precisely known. This can be a major asset to reconstructing paleo geographies and the nature of remnants of fossil ocean basins.
Perhaps the most significant finding, however, is that the crustal section drilled is not matched by that in any known ophiolite. Some of its attributes, including the lack of well developed layering, the presence of small 100- to 500-m intrusions, are similar to structural characteristics from ophiolites formed in slow-spreading environments, such as the Trinity ophiolite or the Josephine ophiolite. However, several of the major features of the section have not been described in these ophiolites. These include: (1) the occurrence of innumerable discrete large and small sheared oxide-rich gabbros intruding undeformed olivine gabbro, and (2) that, strikingly, these decrease downward in abundance through the section; (3) a baseline igneous stratigraphy of successive small "isotropic" olivine gabbro intrusions, the chemistry of which becomes less primitive with depth; (4) synkinematic igneous differentiation of the section, involving intrusion of late iron-rich melts from below into the olivine gabbros at the top of the section along faults and shear zones leading to a Fe-Ti-rich upper section and a lower depleted section; (5) abundant crystal-plastic and brittle-ductile deformation at the top of the section that decreases downward rather than increases as has been described repeatedly in ophiolites.
Together with the well known geochemical affinities of ophiolites for the arc-environment, these "slow-spreading ophiolites" are not good analogs for the Hole 735B section. Nor are "fast-spreading ophiolites" and ocean crust good analogs. The upper gabbros cored during Leg 147 in Hess Deep, and the gabbros of ophiolites inferred to have originated at fast-spreading centers, such as the Oman ophiolite, do not have the extensive crystal-plastic fabrics of the rocks from Hole 735B. The lack of well defined, planar, and continuous igneous layering at Hole 735B also contrasts with the presence of layered gabbros in the lower two-thirds of the gabbros at Oman.
Although these new observations from Leg 176 are generally consistent with the paradigm of the spreading-rate dependence of crustal accretion at ocean ridges, particularly when combined with the recent results of ODP drilling at Hess Deep, they also demonstrate that ophiolite complexes do not fully characterize the evolution of the ocean crust. Penetration and recovery rates at Hole 735B were constant and very high throughout both Legs 118 and 176. Drilling stopped because of a pipe failure during unanticipated weather, not because of hole conditions. Thus, the results of Leg 176 demonstrate the need for future ocean drilling in the lower ocean crust, set the stage for penetrating it to the mantle, and confirm that this can be done using our current drilling platform.
The Nature of Seismic Layer 3
The second major objective of deepening Hole 735B was to test whether
the Moho is coincident with the crust/mantle boundary. Based on an
ocean-bottom seismometer (OBS) refraction study, Muller et al. (1997)
found that the "seismic" crustal thickness north and south of Atlantis
Bank, on which Site 735 is located, is 4 km with a normal-thickness Layer
2 of about 2 km (Fig. 7). These are consistent with crustal thicknesses
predicted and observed for ultra-slow-spreading ridges, including the
Southwest Indian Ridge (Reid and Jackson, 1981; Jackson et al., 1982;
Bown and White, 1994). Beneath Atlantis Bank, however, the projected
depth to Moho is 5 ±1 km, although both drilling and dredge results, to
date, show that the bank exposes only gabbroic rock. Compositions of
basalts dredged from the conjugate site to the north, exposing crust of the
same age as Atlantis Bank, suggest a crustal thickness of 3 ± 1 km (Muller
et al., 1997). Thus, Muller et al. (1997) agree with the predictions based
on sparse geologic evidence (Dick et al., 1991b; Dick, 1996) that the crust
is thin beneath Atlantis Bank and that seismic Layer 3 at that location is
largely composed of partially serpentinized peridotite, with the base of
that layer being an alteration front. We hoped that deepening Hole 735B to
1500+ m would penetrate the crust/mantle boundary and penetrate
partially serpentinized peridotite far above the Moho. Obviously the plus
sign was important here. If this prediction eventually proves true, then
the drilling might demonstrate whether, in at least one place, the Moho is
an alteration front rather than the crust/mantle boundary. This continues
to be a major objective for the drilling program, because, if true, it puts
in question the universal use of this seismic boundary for estimating
crustal thickness in the ocean basins.
Stratigraphy of the Uppermost Mantle
A third major objective of Leg 176 was deepening Hole 735B well below
the crust/mantle boundary to document the shallow mantle stratigraphy
beneath the crust to as great a depth as possible. This would have allowed
determination of the pattern and nature of mantle flow immediately
beneath the ridge axis, the extent of mantle melting beneath a ridge
segment, the mechanism of mantle melt transport to the base of the crust,
and the nature of the primary magmas. The present setting of Hole 735B,
18 km from the trace of the Atlantis II Transform and within the center of
a major tectonic horst block, is an ideal place to observe this stratigraphy
with minimum overprinting by late faulting.
Despite reaching our goal of deepening Hole 735B to a target depth of 1.5 km, we did not penetrate the crust/mantle boundary. Several different factors, including a very high mantle Bouguer Anomaly, several vertical seismic profile (VSP) reflectors immediately beneath the bottom of the hole, and very primitive troctolites found at the base of the Leg 118 section, initially suggested that the crust/mantle boundary might have been encountered immediately beneath the old hole. Instead, we encountered something quite different, drilling into less-primitive gabbros immediately below 500 meters below seafloor (mbsf), and from that point going down through a succession of small gabbro intrusions. With depth, there were a number of systematic changes in the degree of alteration, nature of deformation, and rock chemistry that indicate that the section is not seriously disrupted or imbricated. At the bottom of the hole during Leg 176, we encountered rapidly increasing olivine contents and coarse troctolitic gabbros, which again have raised hopes that we were near a major change in lithology&151perhaps;the crust/mantle boundary. But this is purely hope and speculation at this time. The crust/mantle boundary could be a few hundred meters below the present bottom of the hole, or several kilometers. Because the overall stratigraphy of the lower ocean crust at this location has proved not to have a direct ophiolite analog, we have entered uncharted depths in the Earth.