Introduction
The AAD (Fig. 1) is a unique region within the global mid-ocean spreading system. It encompasses the deepest (4-5 km) region of the global mid-oceanic spreading system. Its anomalous depth reflects the presence of both unusually cold underlying mantle and thin crust. Despite a uniform spreading rate, the eastern boundary of the AAD coincides with an abrupt morphologic change from an axial ridge with smooth abyssal topography off-axis (characteristics usually associated with fast-spreading centers) to deep axial valleys with rough off-axis topography (characteristics usually associated with slow spreading). Other anomalous characteristics of the AAD include a pattern of relatively short axial segments separated by long transforms with alternating offset directions, extremely thin oceanic crust, high upper mantle seismic wave velocities, and an intermittent asymmetric spreading history (Weissel and Hayes, 1971, 1974; Forsyth et al., 1987; Marks et al., 1990; Sempéré et al., 1991; Palmer et al., 1993; West et al., 1994, 1997; Christie et al., 1998). Multiple episodes of ridge propagation from both east and west toward the AAD suggest that the upper mantle is converging toward this region (Vogt et al., 1984, West and Lin, unpubl. data). Indeed, from recent numerical model studies, significant subaxial mantle flow converging on the AAD appears to be an inevitable consequence of gradients in upper mantle temperature around the AAD. Finally, the morphological contrasts across the eastern boundary of the AAD are paralleled by distinct contrasts in the nature and variability of axial lavas, reflecting fundamental differences in magma supply because of strong contrasts in the thermal regime of the spreading center.
Within the easternmost AAD, there is a distinct discontinuity in the Sr, Nd, and Pb isotopic signatures of axial lavas that marks the boundary between Indian Ocean and Pacific Ocean mid-ocean ridge basalt (MORB) mantle provinces (Klein et al., 1988; Pyle et al., 1990; 1992). The boundary itself is remarkably sharp, although there is a gradation within the Pacific region toward Indian Ocean characteristics within 50-100 km of the boundary (Fig. 2). At zero-age seafloor, the boundary is located within 20-30 km of the ~126°E transform—the western boundary of the easternmost AAD spreading segment. The boundary has migrated westward across this segment during the last 3-4 m.y. (Pyle et al., 1990, 1992; Lanyon et al., 1995, Christie et al., 1998) (Figs. 3, 4).
Although such a sharp boundary between ocean-basin-scale upper-mantle isotopic domains is unique along the global mid-ocean ridge system, its long-term relationship to the remarkable geophysical, morphological, and petrological features of the AAD is unclear. The AAD is a long-lived major tectonic feature. Its defining characteristic is its unusually deep bathymetry, which stretches across the ocean floor from the Australian to the Antarctic continental margins. The trend of this depth anomaly forms a shallow west-pointing V-shape cutting across the major fracture zones that currently define the eastern AAD segments (Figs. 1, 4). This V-shape implies that the depth anomaly has migrated westward at a long-term rate of ~15 mm/yr (Marks et al., 1991), which is much slower than the recent migration rate of the isotopic boundary discussed above. The depth anomaly may, in fact, have existed well before continental rifting began ~100 m.y. The presence of restricted sedimentary basins on both continents suggests that precursors of the present AAD may have existed for as long as 300 m.y. (Veevers, 1982; Mutter et al., 1985).
Possible long-term relationships between the isotopic boundary and the morphologically defined AAD fall into two distinct classes, schematically illustrated in Figure 4. Either the recent isotopic boundary migration is simply a localized (~100 km) perturbation of a geochemical feature that has been associated with the eastern boundary of the AAD since the basin opened, or the migration is a long-lived phenomenon that has only recently brought Pacific mantle beneath the AAD. In the first case, the boundary could be related either to the depth anomaly or to the eastern bounding transform, but not to both in the long term. In the second case, the isotopic boundary has only recently arrived beneath the AAD. Although the latter possibility may initially seem fortuitous, it has been independently suggested that Pacific mantle has migrated westward into the region since 40-50 Ma, when separation of the South Tasman Rise from Antarctica first allowed upper mantle flow from the Pacific to the Indian Ocean basin (Alvarez, 1982, 1990). Indian and transitional isotopic signatures from altered ~38- and ~45-Ma basalts dredged to the north and east of the AAD (Lanyon et al., 1995) and from 60- to 69-Ma Deep Sea Drilling Project (DSDP) basalts that were drilled close to Tasmania (Pyle et al., 1992) provide limited support for this hypothesis. Recent off-axis sampling in Zone A (Christie et al., 1998) constrains any such boundary to lie within the shaded region of Figure 4 and perhaps requires a hiatus of at least 3 m.y. between the first arrival of Pacific mantle at the eastern boundary of the AAD and its initial penetration into the AAD proper (West and Christie, 1997; Christie et al., 1998).
The Nature of the Indian Ocean MORB Mantle Province
The mantle source for Indian Ocean MORB is distinct from that of the Pacific Ocean MORB in having distinctly lower 206Pb/204Pb and 208Pb/204Pb and higher 87Sr/86Sr, as well as systematically lower 207Pb/204Pb and 143Nd/144Nd (Figs. 2, 3). The sharpness of the Indian/Pacific boundary, as expressed in the seafloor lavas, suggests that Indian MORB mantle presently abuts Pacific MORB mantle beneath the AAD, with little or no intermingling. In contrast, along the Southwest Indian Ridge, there is a much more gradational transition from Indian- to Atlantic-type mantle (Mahoney et al., 1992).
The distinctive characteristics of Indian MORB mantle have been variously attributed to the widespread dispersal throughout an otherwise "typical" depleted upper mantle of material with distinctive isotopic characteristics derived from one or more of the following: (1) Indian Ocean hot spot sources, especially the large long-lived Kerguelen mantle plume, (2) lower continental lithosphere derived from the breakup of Gondwanaland, and/or (3) convectively recycled subducted altered oceanic crust (e.g., Subbarao and Hedge, 1973; Hedge et al., 1973; Dupré and Allègre, 1983; Hamelin et al., 1985; Hamelin and Allègre, 1985; Hart, 1984; Michard et al., 1986; Price et al., 1986; Dosso et al., 1988; Klein et al., 1988; Mahoney et al., 1989). The Indian MORB isotopic signature has also been attributed to the interaction of Gondwana continental lithosphere with the Kerguelen mantle plume before India rifted from Australia (Storey et al., 1988; Mahoney et al., 1989, 1992).
Away from the spreading centers, the extent of the Indian MORB mantle is only poorly known. In the region of interest for Leg 187, Pyle et al. (1992) analyzed all available drilled material. They showed that Indian mantle has been present at 110°E, to the east of Kerguelen, since at least 30-40 Ma, and that it may have been present to the east of the AAD before 39 Ma (Pyle et al., 1992; Lanyon et al., 1995). No basalts of Indian affinity have been reported east of the South Tasman Rise at any age, and none younger than 30 Ma are known anywhere east of the AAD. In addition, all samples analyzed so far from recent sampling of the SEIR west of the AAD are clearly of Indian type (L. Hall, J. Mahoney, pers comm., 1998).
The dispersion of Indian mantle and its areal extent may be controlled by one or more of the following:
1. Flattening of the heads of large mantle plumes (~2000 km) (Mahoney et al., 1992);
2. Global-scale upper mantle convection (Hamelin and Allègre, 1985) and, more specifically, advection by temperature gradient-driven mantle flow within the ocean basin (West et al., 1997);
3. Isolation of the upper mantle by the deep roots of the surrounding Gondwana continents (Alvarez, 1982, 1990); and
4. Restriction of this upper mantle province to the limits of Archean subcontinental lithosphere beneath the Gondwana continents (Klein et al., 1988).
Regardless of its origin and evolution, the nature and behavior of this isolated reservoir can be better understood through a better definition of the configuration and, hence, the dynamics of its eastern boundary. Because this boundary is so sharply defined and uncontaminated by hot spots or other nearby perturbations and because the plate motions between Australia and Antarctica are uncomplicated and well known, simple testable predictions can be made for a broad range of hypotheses.
The Origin and Evolution of the Isotopic Boundary
The most direct objective of this proposal is to define, as closely as possible, the off-axis configuration of the Indian/Pacific mantle isotopic boundary. In addition to its importance as a "local" phenomenon, an improved understanding of this boundary is important for a broader general understanding of the oceanic mantle. In investigating the origins of the AAD and the isotopic boundary, we are also investigating the importance of variations in geochemistry, isotopic makeup, temperature, and other physical characteristics of the oceanic upper mantle in general. Improved knowledge of the distribution of these chemical and physical characteristics in space and time will lead to a better understanding of the dynamics of the oceanic mantle and of its interaction with the magmatic processes of the mid-ocean ridge system.
Three possible end-member configurations of the isotopic boundary on the Southern Ocean seafloor are illustrated in Figure 4. In the simplest configuration, the isotopic boundary has always been associated with the eastern boundary of the AAD and therefore follows a flow line oriented approximately north-south. Small-scale (~100 km) perturbations in the east-west position of the Indian/Pacific MORB boundary would be consistent with the apparent westward migration of the boundary along segment B5 in the eastern AAD during the last 4 m.y. In the second case, the boundary is associated with the depth anomaly and follows its trace off-axis. The V-shaped cofiguration of this trace requires that it has moved westward at ~15 mm/yr (Marks et al., 1991); whereas, the recent migration rate of the isotopic boundary is 25-40 mm/yr (Pyle et al., 1992, Christie et al., 1998), again requiring small-scale east-west fluctuations in the boundary position to be superimposed on the more gradual (~15 mm/yr) westward motion. In the third case, the isotopic boundary is produced by steady westward migration of Pacific mantle since rifting of the South Tasman Rise. In this case, a reasonable rate for Pacific mantle inflow can be calculated from the assumption that a continental barrier to mantle flow was removed at ~40 Ma, when circum-Antarctic ocean circulation was established south of Tasmania (Royer and Sandwell, 1989; Mutter et al., 1985). This rate is comparable to the recent migration rate of the boundary within the AAD and to the propagation rates (which likely reflect mantle flow; West and Lin, unpubl. data) of three westward-propagating rifts along the SEIR east of the AAD. This rate is a long-term average, however, and systematic variations in the along-axis migration rate could be expected with the opening of the ocean basin (West et al., 1997).
Subsurface Biosphere
Recent findings have extended the biosphere to include microbial life in deep subsurface volcanic regions of the ocean floor (Thorseth et al., 1995; Furnes et al., 1996; Fisk et al., 1998; Torsvik et al., 1998). Much attention has been recently focused toward the existence of microbes living on and contributing to the alteration of basaltic glass in lavas from the upper part of the oceanic crust (Thorseth et al., 1995; Furnes et al., 1996; Fisk et al., 1998; Torsvik et al.,
1998). The first recognized evidence of this phenomena was from textures in basaltic glass from Iceland (Thorseth et al., 1992). Similar textures were later found in basaltic glass from ODP Hole 896A at the Costa Rica Rift, and the microbial contribution to the alteration history was supported by the presence of DNA along the assumed biogenic alteration fronts (Thorseth et al., 1995; Furnes et al., 1996; Giovannoni et al., 1996). Microbes have recently been documented to inhabit internal fracture surfaces of basaltic glass that specifically were sampled for microbiologic studies during Mir submersible dives to the Knipovich Ridge (Thorseth et al. 1999). The presence of dissolution textures underneath many microbes, and manganese and iron precipitates next to microbes, suggests that microbial activity does play an active role in the low-temperature alteration of ocean-floor basalts.
The planned sampling of ocean-floor basalts that range in age from 7 to 30 Ma provides an opportunity to study how microbial alteration progresses with time. It also allows us to isolate microbes in ODP samples taken under specific conditions where the effect of drilling related contamination may be evaluated.