Beginning with the early studies of Schlich (1975) and Houtz et al. (1977), and as summarized by Munschy and Schlich (1987), the plateau has been divided into two distinct domains. The Northern Kerguelen Plateau (NKP), ~46° to 54°S, has shallow water depths, <1000 m, and basement elevations 3-4 km above the adjacent seafloor with the maximum elevations forming the Kerguelen, Heard, and McDonald Islands (Fig. 2). Broken Ridge and the Northern Kerguelen Plateau are conjugate Early Cretaceous provinces (Fig. 1) that were separated by seafloor spreading along the Southeast Indian Ridge (SEIR) during the Eocene (Mutter and Cande, 1983). The Southern Kerguelen Plateau (SKP) is generally characterized by deeper water, 1500 to 2500 m, and it is tectonically more complex (Figs. 2, 4). It is characterized by several large basement uplifts and was affected by multiple stages of normal faulting, graben formation, and strike-slip faulting (e.g., Coffin et al., 1986; Fritsch et al., 1992; Rotstein et al., 1992; Royer and Coffin, 1992; Könnecke and Coffin, 1994; Angoulvant-Coulon and Schlich, 1994). On the basis of gravity data, seismic results, and the study of sedimentary sections (Fröhlich and Wicquart, 1989), Coffin and Eldholm (1994) infer a break between the Mesozoic and Cenozoic parts of the plateau immediately south of the Kerguelen Archipelago and north of a major sedimentary basin (Fig. 4).
Crustal Structure
Based on multichannel seismic reflection data, Schaming and Rotstein
(1990) suggested that basaltic crust forms the northern and southern
parts of the Kerguelen Plateau; however, volcanic morphology is largely
absent in the SKP because of the subaerial erosion that preceded
subsidence. Numerous dipping basement reflectors that are interpreted as
flood basalt constructions have been identified in the crust of the
Southern Kerguelen Plateau (Coffin et al., 1990; Schaming and Rotstein,
1990; Schlich et al., 1993). Ocean-bottom-seismometer wide-angle
reflection and refraction experiments have been undertaken recently on
both the northern and southern sectors of the plateau (Charvis et al.,
1993, 1995; Operto and Charvis, 1995, 1996). On the northern plateau,
igneous crust can be divided into two parts: (1) an upper Layer 2 with an
average thickness of 5.5 km and a velocity gradient ranging from 4.3 to 6.1
km/s, and (2) a lower layer ~17 km thick with velocities from 6.6 to 7.4
km/s (Charvis et al., 1995; Recq et al., 1994). This crustal structure
differs significantly from that below the Kerguelen Archipelago where the
upper igneous crust is thicker, ~10 km, and the lower crust is thinner, 7-9
km (Recq et al., 1990). These observations, in conjunction with geological
sampling and tectonic reconstructions, are used to infer that the NKP from
the Kerguelen Archipelago northwestward is younger than 42 Ma (Royer
and Coffin, 1992; Munschy et al., 1992; Charvis et al., 1993, 1995). On the
southern plateau, the igneous crust can be divided into three layers: (1) an
upper crustal layer ~5.3 km thick with velocities ranging from 3.8 to 6.5
km/s; (2) a lower crustal layer ~11 km thick with velocities of 6.6 to 6.9
km/s; and (3) a 4- to 6-km-thick transition zone at the base of the crust
characterized by an average velocity of 6.7 km/s (Recq et al., 1994; Operto
and Charvis, 1995, 1996). This low velocity, seismically reflective
transition zone at the crust-mantle interface has not been imaged on the
NKP, and it is the basis for the hypothesis that parts of the SKP are
fragments of a volcanic passive margin, similar to the Rockall Plateau in
the North Atlantic Volcanic Province (Schlich et al., 1993; Operto and
Charvis, 1995, 1996).
Previous Sampling of Igneous Basement
What was learned from previous sampling (ODP Legs 119 and 120 and
dredging) of the Kerguelen Plateau/Broken Ridge LIP? Based in large part
on ODP-related studies, we can now conclude that decompression melting
of the Kerguelen Plume was a major magma source for the Kerguelen
Plateau/Broken Ridge system. Although sampling and age dating are
grossly insufficient, large volumes of subaerially erupted tholeiitic
basalt magma formed much of the central and southern plateau during the
interval from ~115 to 85 Ma. Specifically, constraints from the ages of
overlying sediments combined with K-Ar and
40Ar/39Ar dating of a dredged basalt (Leclaire
et al., 1987), and basalts from ODP Leg 120 (Whitechurch et al., 1992) led
to an inferred emplacement age of 110-120 Ma for the central portion of
the Kerguelen Plateau. The age data base has been recently augmented by
the 40Ar/39Ar study of Pringle et al. (1994) and
Storey et al. (1996) who report ages ranging from 108.6 Ma to 112.7 Ma
for basement basalts from ODP Sites 738, 749, and 750; therefore, south
of 57°S the uppermost Kerguelen Plateau is ~110 Ma. In contrast, basalts
from Site 747 on the central part of the plateau are much younger, ~85 Ma.
This age is similar to the 83-88 Ma age for lavas from Broken Ridge
Dredge Sites 8 and 10 (Fig. 3; Duncan, 1991), which is consistent with the
pre-rifting position of Site 747 relative to these Broken Ridge dredge
sites. Also, piston coring of sediments on the northeast flank of the
plateau between the Kerguelen Archipelago and Heard Island recovered
upper Cretaceous cherts and calcareous oozes of probable Santonian age
(Leclaire et al., 1987). In summary, we have very few high quality age data
for the 2.3 x 106 km2 (equivalent to ~8 Icelandic
plateaus) of the Kerguelen Plateau/Broken Ridge LIP. Nevertheless, it is
evident that large magma volumes erupted over short time intervals,
possibly as two or even three pulses: the SKP at ~110 Ma, the central KP
and Broken Ridge at ~85 Ma and much of the Kerguelen Archipelago and
perhaps the northernmost portion of the Kerguelen Plateau at ~40-23 Ma
(Nicolaysen et al., 1996). Sampling by drilling at other sites throughout
the plateau is required to determine if formation of this LIP was truly
episodic or if there was a continuous south to north decrease in the age of
volcanism.
Although the Kerguelen Plateau is a volcanic construction formed in a young oceanic basin (Royer and Coffin, 1992; Munschy et al., 1994), evidence is equivocal as to whether it was emplaced at a spreading center (e.g., Iceland) or off-ridge (e.g., Hawaii; Coffin and Gahagan, 1995). Oxidized flow tops and the vesicularity of lava flows at ODP Sites 738 and 747, however, are consistent with eruption in a subaerial environment. Moreover, at ODP Site 750 the basement lavas are overlain by nonmarine, organic-rich sediments containing up to 5-cm pieces of charcoal. Coffin (1992) concluded that the drill sites in the SKP had long, >10 Ma, histories of subaerial volcanism and erosion followed by subsidence caused by cooling. Higher temperature metamorphism, based on zeolite mineralogy, at Site 749 and then at Sites 747 and 750 may indicate erosion to deeper levels at Site 749 (Sevigny et al., 1992).
The islands on the NKP are dominantly formed of <40-Ma transitional and alkaline lavas. In contrast, dredging along the 77E graben of the Kerguelen Plateau and from Broken Ridge (Figs. 1, 4) recovered basaltic rocks whose compositions are similar to ocean island tholeiites; they clearly are not typical mid-ocean-ridge basalt (MORB) (e.g., Davies et al., 1989; Weis et al., 1989; Mahoney et al., 1995). Tholeiitic basalts also form the igneous basement of the Kerguelen Plateau at drill Sites 738, 747, 749, and 750 (Fig. 5). Except for an alkaline basalt flow 200 m above basement at Site 748, the long-term volcanism, ~115 to 38 Ma, associated with the Kerguelen Plume, i.e. the Kerguelen Plateau and Ninetyeast Ridge, are exclusively tholeiitic (e.g., Frey et al., 1991; Saunders et al., 1991; Storey et al., 1992; Frey and Weis, 1995). The significance of this result is that tholeiitic compositions are derived by relatively high extents of melting, (e.g., Kent and McKenzie, 1994), and the inference is that the Kerguelen plume was a high-flux magma source for a long time. However, to date the MgO and olivine-rich melts expected from the melting of large amounts of high-temperature plumes (Storey et al., 1991) have not been recovered. In fact, in contrast to the tholeiitic lavas forming the Hawaiian shields, there is no evidence for melt segregation at relatively high pressures within the garnet stability field (Frey et al., 1991).
Most of the lavas from the Kerguelen Plateau and Broken Ridge have Sr and Nd isotopic ratios which range from ratios typical of enriched MORB from the SEIR to ratios proposed for the Kerguelen Plume (Fig. 5). In Pb-Pb isotopic ratios most of the lavas from the Kerguelen Plateau define an elongated field that is subparallel to that for SEIR MORB; however, like lavas forming the Kerguelen Archipelago, the Kerguelen Plateau lavas are offset from the MORB field to higher 208Pb/204Pb at a given 206Pb/204Pb composition (Fig. 5). These isotopic data have been interpreted as resulting from mixing of the Kerguelen Plume with entrained depleted asthenosphere (e.g., Weis et al., 1992). In contrast, basalts from ODP Site 738 on the southern extremity of the Kerguelen plateau and Dredge 8 from eastern Broken Ridge (Fig. 1) have atypical geochemical characteristics for oceanic lavas. These lavas have very high 87Sr/86Sr, low 143Nd/144Nd, and high 207Pb/204Pb ratios, which accompany relatively low 206Pb/204Pb compositions (Fig. 5). Although the sampling density is limited, Mahoney et al. (1995) show that lavas from plateau locations closest to continental margins, e.g., Site 738 in the far south, Dredge 8 from eastern Broken Ridge, and lavas from the Naturaliste Plateau (Fig. 1), have the most extreme isotopic characteristics (e.g., 87Sr/86Sr >0.7090), which are accompanied by relative depletions in Nb and Ta and relatively high 207Pb/204Pb ratios. They conclude that these geochemical features arose from a continental lithosphere component (e.g., Storey et al., 1989) that contributed to magmatism near the edges of the Kerguelen Plateau/Broken Ridge system. The geochemical evidence for this continental component is consistent with geophysical evidence for the southernmost part of the plateau consisting of a passive margin fragment (Schlich et al., 1993; Operto and Charvis, 1995).
Several lines of evidence support the interpretation that the Kerguelen Plume has been a long-term source of magma for major bathymetric features in the eastern Indian Ocean. For example, the systematic north-south age progression on the Ninetyeast Ridge is consistent with a hot-spot track formed as the Indian plate migrated northwards over the Kerguelen Plume (Duncan, 1991). Also the isotopic similarities between lavas from the Ninetyeast Ridge, with the younger lavas forming the Kerguelen Archipelago and Heard Island and the older lavas forming the Kerguelen Plateau and Broken Ridge (Fig. 5), indicate that the Kerguelan Plume played an important role (Weis et al., 1992; Frey and Weis, 1995, 1996). However, a continental lithosphere source component has been recognized in some lavas from the SKP and eastern Broken Ridge (Mahoney et al., 1995). Also wide-angle seismic data collected by ocean bottom seismometers in the Raggat Basin of the SKP have defined a reflective zone at the lower crust-mantle boundary that was interpreted to be stretched continental lithosphere (Recq et al., 1994; Operto and Charvis, 1995, 1996). The present geochemical data set shows that a continental lithosphere component is obvious in lavas at only two sites (Dredge 8 on BR and Site 738 in the SKP, Fig. 5). There is no evidence for a continental component in lavas from the Central Kerguelen Plateau, the Ninetyeast Ridge, and the Kerguelen Archipelago (Frey et al., 1991). Determination of the spatial and temporal role of this lithosphere component is required to evaluate if this continental component is a piece of Gondwana lithosphere that was incorporated into the plume.
Because the isotopic characteristics of plume, asthenosphere, and lithosphere sources are usually quite different, the temporal geochemical variations in stratigraphic sequences of lavas can be used to determine the relative roles of plume, asthenosphere, and lithosphere sources in plume-related volcanism; in particular, the proportion of these sources changes systematically with time and location (e.g., Chen and Frey, 1985; Gautier et al., 1990; White et al., 1993; Peng and Mahoney, 1995), thereby providing an understanding of how plumes "work." In addition to answering questions about plume-lithosphere interactions, geochemical data for plateau lavas will define the role of depleted asthenosphere in creating this plateau. A MORB-related asthenosphere is apparent in some of the Ninetyeast Ridge drill sites (e.g., as reflected by the Sr and Nd isotopic ratios of lavas from Site 756; Weis and Frey, 1991) and is an expected consequence of the plume being close to a spreading ridge axis during formation of the Ninetyeast Ridge. Transects of relatively shallow basement holes (200 m) in the Kerguelen Plateau can be used to define spatial and short term variability during the waning phase of plateau volcanism. A surprising result of the shallow penetrations of the Kerguelen Plateau is that sampling of 35-50 m of igneous basement at several plateau sites shows that lavas at each site have a suite of distinctive geochemical characteristics: each site has a distinctive combination of Sr and Nd isotopic ratios (Fig. 5). Does this heterogeneity reflect the spatial heterogeneities in a plume or localized differences in mixing proportions of components derived from asthenosphere, plume, and slivers of continental lithosphere? Interpretation requires a knowledge of temporal variations in geochemical characteristics at several locations.
KIP-2B
This site is on the NKP, which has never been sampled. A site north of the
Kerguelen Archipelago is required to establish the age, composition, and
emplacement environment of the NKP (presumed to be Cenozoic in age), as
well as to test plate reconstructions. Site KIP 2B is in shallow water
(~150 m) and efforts are being made to identify a suitable deep-water
site on the NKP.
KIP-3A
Several distinct, roughly circular bathymetric and gravity highs form a
linear trend between the Kerguelen Archipelago and Heard Island (Figs. 3,
4), the two recently active volcanic islands in the NKP. This site was
selected because it is on one of several aligned gravity highs that may
reflect a hotspot track between Heard and Kerguelen Islands.
KIP-6B
This site is on Elan Bank, which is a prominent east-west oriented
bathymetric and gravity high west of the Central Kerguelen Plateau (Figs.
2, 4). This major morphotectonic component of the Kerguelen Plateau
province is unsampled; the site will constrain its age, composition,
emplacement environment, and uplift-subsidence history.
KIP-7A
This site in the Central Kerguelen Plateau is about 100 km north of ODP
Site 747. The determined age of basalts at Site 747 is ~85 Ma and
represents the only sampled basement of Late Cretaceous age found on the
Kerguelen Plateau. Drilling at Site KIP-7A will evaluate the areal extent
of Late Cretaceous crust. In addition, the sedimentary section will help
constrain the post-emplacement tectonic history.
KIP-12A
This site on the Southern Kerguelen Plateau is located on a bathymetric
and gravity high (Figs. 2, 4). The location was selected to evaluate the
areal extent of continental crustal contamination north of ODP Site 738
(Figs. 2, 5), as well as the age of this part of the plateau. Several hundred
meters of sediment will help to constrain the uplift/subsidence history.
KIP-9A
This site on the Eastern Broken Ridge will provide the first in situ
basement samples from the entire Broken Ridge (Fig. 3). Although the final
location is waiting on the processing and subsequent examination of
Conrad 2708 seismic data, it will be located near Dredge 8, whose lavas
are similar in age (88 Ma) to lavas from ODP Site 747. Unlike lavas from
the central part of the Kerguelen plateau, Dredge 8 lavas contain a
continental lithosphere component (Fig. 5). This site will define the
spatial extent of this component.
The Formation MicroScanner (FMS) provides a detailed resistivity image of the borehole wall. It has previously been used successfully in a volcanic setting on Leg 152 (Cambray, in press; Planke and Cambray, in press), and will be recorded in all basement holes on Leg 183. The log provides a detailed volcano stratigraphy and is particularly useful to image altered and fractured zones which may be poorly sampled during coring. It also has potential to provide important structural and stress field information, and it can be used for core-log integration.
Seismic reflection data are essential to extrapolate the drilling results away from and between the drill sites. Synthetic seismograms constructed from the velocity and density logs provide a link between the core and the reflection data. Reliable synthetic seismograms require calibration with downhole seismic data. These calibration data will be obtained by using a well sonic tool (WST) in holes with a sediment penetration of more than 500 m.
Other speciality tools will only be run at selected sites. The dipole shear
imager (DSI) tool can provide shear-wave velocity information in both the
relatively slow lava flow margins and relatively fast flow interiors. Such
data are important as inputs for modeling seismic wave propagation in
volcanic sequences. The Azimuthal Resistivity Imager (ARI) is a
high-resolution deep resistivity imaging tool. The data are useful for
obtaining formation dip and thin-bed resistivity as the measurements are
not influenced by drilling-induced fractures. The borehole televiewer
(BHTV) provides an acoustic image of the borehole wall, which is useful
for structural and stress field analysis (e.g., Paillet and Kim, 1987;
Castillo and O'Neill, 1992). The interpretation of geochemical stratigraphy
including alteration zones is facilitated by the geochemical tool (GLT)
(e.g., Pratson et al., 1992). At this time we note the possibility of
utilizing the GLT, but do not include it as definite part of the Logging Plan
(Table 1).
Table of Contents