STUDY AREA

Physical Description

The Kerguelen Plateau is a broad bathymetric high in the southern Indian Ocean surrounded by deep ocean basins—to the northeast by the Australian-Antarctic Basin, to the south by the 3500-m-deep Princess Elizabeth Trough, to the southwest by the Enderby Basin, and to the northwest by the Crozet Basin (Figs. F1, F3). The plateau stretches ~2300 km between 46°S and 64°S in a southeast-trending direction toward the Antarctic continental margin. It is between 200 and 600 km wide and stands 2-4 km above the adjacent ocean basins. Variable age oceanic crust abuts the Kerguelen Plateau. As summarized by Schlich and Wise (1992), the oldest magnetic anomalies range from Chron C11n.2n (o) (30.1 Ma), in the northeast, to Chron C18n.2n (o) (40.1 Ma) off the central part of the eastern plateau (we use the geomagnetic polarity time scale of Cande and Kent, 1995). Farther south, the eastern flank of the Southern Kerguelen Plateau is bounded by the Labuan Basin. Basement of the Labuan Basin has not been sampled by drilling, but its structure resembles that of the main Kerguelen Plateau (Rotstein et al., 1991; Munschy et al., 1992). To the northwest, magnetic anomaly sequences from Chrons C23 to C34 have been identified in the Crozet Basin, but to the southwest no convincing anomalies have been identified in the Enderby Basin, although Mesozoic anomalies have been suggested (Li, 1988; Nogi et al., 1996). An Early Cretaceous age for the Enderby Basin is assumed in most plate reconstructions (e.g., Royer and Coffin, 1992).

Beginning with early studies (Schlich, 1975; Houtz et al., 1977), the Kerguelen Plateau province has been divided into distinct domains. Coffin et al. (1986) and Könnecke et al. (1998) recognize five domains: northern, central, and southern Kerguelen Plateau; Elan Bank; and the Labuan Basin (Figs. F3, F4). The northern Kerguelen Plateau (NKP), ~46°S to 50°S, has shallow water depths (<1000 m) and basement elevations 3000-4000 m above adjacent seafloor, with maximum elevations forming the Kerguelen Archipelago. A lack of rocks older than ~40 Ma from the Kerguelen Archipelago (Nicolaysen et al., 1996, in press), as well as plate reconstructions (Royer and Sandwell, 1989; Royer and Coffin, 1992), suggests that the age of the NKP is 40 Ma, whereas the central and southern domains of the submarine Kerguelen Plateau appear to be of Cretaceous age (Pringle et al. 1994; Storey et al., 1996). However, the basement of the submarine NKP had not been sampled before Leg 183. The central Kerguelen Plateau (CKP), ~50°S to 55°S, is also relatively shallow, contains a major sedimentary basin (Kerguelen-Heard Basin), and includes the volcanically active Heard and McDonald Islands. Broken Ridge and the CKP are conjugate Late Cretaceous provinces (Fig. F1) that were separated by seafloor spreading along the SEIR during Eocene time (Mutter and Cande, 1983).

The southern Kerguelen Plateau (SKP) apparently formed in Early Cretaceous time (Pringle et al., 1994; Storey et al., 1996). Relative to the NKP, water depths are greater (1500 to 2500 m), and it is tectonically more complex (Figs. F3, F4). There are several large basement uplifts and evidence for 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; Angoulvant-Coulon and Schlich, 1994; Könnecke and Coffin, 1994; Gladczenko et al., 1997). Elan Bank, a salient extending westward from the boundary between the CKP and SKP, has water depths from <1000 to 2000 m. Before Leg 183, basement had not been sampled from Elan Bank and its age was, therefore, unknown. The Labuan Basin, which adjoins the CKP and SKP to the east, is a deep (>3500 m), extensively faulted, thickly sedimented (>2 s two-way traveltime in places, or >2000 m, assuming a sediment velocity of 2000 m/s) basin. Dredging of an exposed faulted basement block in the Labuan Basin recovered metamorphic and granitic rock; these rocks have been interpreted as ice-rafted debris (Montigny et al., 1993). Thus, the basin's age and the nature of its crust (i.e., oceanic or continental) remain uncertain.

The CKP was contiguous with Broken Ridge when these domains formed during Cretaceous time (Duncan, 1991; Houtz et al., 1977; Pringle et al., 1994; Storey et al., 1996). Subsequently, at ~40 Ma, Broken Ridge and the CKP began to separate along the westward-propagating SEIR. Broken Ridge, now ~1800 km north of the Kerguelen Plateau, is a narrow and elongated oceanic plateau (100-200 km by ~1000 km at ~2000 m water depth) that trends west-northwest (Figs. F5, F6). It is markedly asymmetric in cross section, dipping gently (<2°) toward the north but with a steeply dipping (>10°) southern face (Fig. F5). This southern flank was uplifted, perhaps more than 2000 m, during the early Tertiary breakup between Broken Ridge and the Kerguelen Plateau (Peirce, Weissel, et al., 1989; Weissel and Karner, 1989).

Crustal Structure

Ocean Drilling Program (ODP) Legs 119 and 120 drilling results (Barron, Larsen, et al., 1989; Schlich, Wise, et al., 1989), dredging data (Leclaire et al., 1987; Davies et al., 1989; Duncan, 1991; Weis et al., 1998a), and multichannel seismic reflection data (Coffin et al., 1990; Schaming and Rotstein, 1990; Schlich et al., 1993) have shown that igneous basement of the Kerguelen Plateau and conjugate Broken Ridge is basaltic. Numerous dipping intrabasement reflections interpreted as flood basalts have been identified in the crust of the CKP and SKP and on Elan Bank (Könnecke et al., 1997). Wide-angle seismic data from the Kerguelen Archipelago on the NKP show an upper igneous crust 8-9.5 km thick and a lower crust 6-9.5 km thick (Charvis et al., 1995; Recq and Charvis, 1986; Recq et al., 1990, 1994). Wide-angle reflection and refraction experiments employing ocean-bottom seismometers have been undertaken recently on both the CKP and SKP (Charvis et al., 1993, 1995; Operto and Charvis, 1995, 1996; Könnecke et al., 1998; Charvis and Operto, 1998, 1999). The crustal structure beneath the Kerguelen Archipelago differs significantly from that of the CKP. Igneous crust of the CKP is 19 to 21 km thick and is composed of three layers. The upper layer is 1.2 to 2.3 km thick, and velocities range from 3.8 to 4.9 km/s. It could be composed of either lava flows or interlayered volcanic and sedimentary beds. The second layer is 2.3 to 3.3 km thick, and velocities increase downward from 4.7 to 6.7 km/s. In the ~17-km-thick lower crust, velocities increase from 6.6 km/s at ~8.0 km depth (near the top of the layer) to 7.4 km/s at the base of the crust, with no internal discontinuity. On the southern plateau, the ~22-km-thick 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 velocities of 6.7 to 6.9 km/s (Operto and Charvis, 1995, 1996). This low-velocity, seismically reflective transition zone at the crust/mantle interface has not been imaged on the NKP or CKP; it is the basis for the hypothesis that parts of the SKP contain fragments of continental crust (Operto and Charvis, 1995; 1996).

Previous Sampling of Igneous Basement: Ages and Geochemical Characteristics

In this section we summarize results of previous sampling (Legs 119 and 120 and dredging) of the Kerguelen Plateau-Broken Ridge LIP. Based in large part on ODP-related studies, there is a consensus that the Kerguelen plume was the major source of magma for constructing the Kerguelen Plateau-Broken Ridge LIP. Although samples and dates from the entire LIP are meager, sampling of the SKP at four spatially diverse locations (Sites 738, 749, and 750, and dredge Site MD48-05; see Figs. F3, F4) shows that the uppermost igneous crust of SKP formed over a relatively short interval at ~110 Ma (K/Ar data from Leclaire et al., 1987; Whitechurch et al., 1992; and ⁴⁰Ar/³⁹Ar data from Pringle et al., 1994; Storey et al., 1996). In contrast, basement basalts from Site 747 on the CKP may be much younger, ~85 Ma (Pringle et al., 1994; Storey et al., 1996). This age is similar to the 83- to 88-Ma age for lavas from Broken Ridge dredge sites 8 and 10 (⁴⁰Ar/³⁹Ar data from Duncan, 1991), which coincide spatially with the prebreakup position of Site 747 (Figs. F1, F2). Also, piston coring of sediments on the northeast flank of the CKP between the Kerguelen Archipelago and Heard Island (MD35-510 in Fig. F4) recovered cherts and calcareous oozes of probable Santonian age (Fröhlich and Wicquart, 1989). In summary, we have very few high-quality age data for the 2.3 × 10⁶ km² (equivalent to approximately eight Icelandic plateaus) of the Kerguelen Plateau-Broken Ridge LIP. These sparse data support the hypothesis that large magma volumes erupted over short time intervals, possibly as two pulses during Cretaceous time—the SKP at ~110 Ma; the CKP, Broken Ridge, and perhaps Elan Bank at ~85 Ma (Fig. F7). In contrast, Cenozoic volcanism (~38 Ma to present) has formed the Kerguelen Archipelago (e.g., Nicolaysen et al., 1996, in press), Heard and McDonald Islands (Clarke et al., 1983; Quilty et al., 1983), and the bathymetric/gravity highs between the Kerguelen Archipelago and Heard Island (Weis et al., 1998a). A major goal of Leg 183 was to drill at other sites throughout the plateau to determine whether formation of this LIP was truly episodic or continuous with a south to north decrease in age of volcanism.

Although the southern and central domains of the Kerguelen Plateau formed in a young oceanic basin (Royer and Coffin, 1992; Munschy et al., 1994), evidence is equivocal as to whether it formed at a spreading center, like Iceland, or off-ridge, like Hawaii (Coffin and Gahagan, 1995). Before Leg 183, several observations had indicated that much of the uppermost basement of the southern and central Kerguelen Plateau erupted in a subaerial environment—specifically, (1) oxidized flow tops and vesicularity of lava flows at Sites 738 and 747; (2) nonmarine, organic-rich sediments containing up to 5-cm pieces of charcoal overlying the basement at Site 750; and (3) claystone overlain by a basalt cobble conglomerate and glauconitic sediment with wood fragments in the lowermost core at Site 748 (Schlich, Wise, et al., 1989). Coffin (1992) concluded that the drill sites in the SKP had long (>10 to 50 m.y.) histories of subaerial volcanism and erosion, followed by subsidence caused by cooling. Zeolite mineralogy of the basaltic basement indicates temperatures of 120°C at Sites 747 and 750, but higher temperatures (225°C) at Site 749, perhaps indicating deeper levels of erosion (Sevigny et al., 1992).

The islands on the Kerguelen Plateau are dominantly formed of <40-Ma transitional and alkaline lavas (Fig. F8) (Weis et al., 1993, 1998b; Barling et al., 1994; Yang et al., 1998; Nicolaysen et al., 1996, in press). Before Leg 183, the only alkaline basalt recovered from the Kerguelen Plateau, Broken Ridge, and Ninetyeast Ridge was a flow ~200 m above basement at Site 748. Tholeiitic basalt of Cretaceous age has been recovered from four dredge and four drill sites on the central and southern Kerguelen Plateau and three dredge sites on Broken Ridge (Figs. F3, F4, F5, F6, F8); seven drill sites on Ninetyeast Ridge have yielded solely tholeiitic basalt ranging from 38 to 82 Ma (Fig. F1). Tholeiitic basalts from each of these sites are geochemically distinct, but as a group, their incompatible element abundances resemble those of ocean-island tholeiitic basalts, rather than typical mid-ocean-ridge basalts (MORBs) (data sources: Kerguelen Plateau and Broken Ridge: Davies et al., 1989; Weis et al., 1989; Storey et al., 1992; Mahoney et al., 1995; Ninetyeast Ridge: Frey et al., 1991; Saunders et al., 1991; Frey and Weis, 1995). We infer that tholeiitic basalt was the dominant magma type produced by the Kerguelen plume from ~110 to 38 Ma during formation of the Kerguelen Plateau, Broken Ridge, and Ninetyeast Ridge. The significance is that tholeiitic basalts are derived from relatively high (>5%) extents of partial melting (Kent and McKenzie, 1994), and the inference is that the Kerguelen plume was a high-flux magma source for a long time (Figs. F2, F7). However, the MgO-rich melts expected from large extents of melting of high-temperature plumes (e.g., Storey et al., 1991) have not been recovered from Cretaceous parts of the Kerguelen Plateau. Picritic (i.e., olivine rich) alkaline lavas of Quaternary age are found on Heard Island (Barling et al., 1994), and transitional picritic lavas (14 to 19 Ma) were recently dredged from one of the bathymetric/gravity highs between the Kerguelen Archipelago and Heard Island (Weis et al., 1998a) (see Figs. F3, F4). All of these picrites are olivine-rich cumulates rather than crystallized MgO-rich melts.

Most lavas from the Kerguelen Plateau and Broken Ridge have Sr and Nd isotope ratios that range from the high ⁸⁷Sr/⁸⁶Sr-low ¹⁴³Nd/¹⁴⁴Nd end of the field for SEIR MORB to the field proposed for the Kerguelen plume (Fig. F9). In Pb-Pb isotope plots, Kerguelen Plateau lavas from Sites 747, 749, and 750 define an elongate field subparallel to that for SEIR MORB (Fig. F10); however, like lavas forming the Kerguelen Archipelago, the Kerguelen Plateau lavas are offset from the MORB field to higher ²⁰⁸Pb/²⁰⁴Pb and ²⁰⁷Pb/²⁰⁴Pb at a given ²⁰⁶Pb/²⁰⁴Pb ratio. In addition, submarine Kerguelen Plateau lavas extend to lower ²⁰⁶Pb/²⁰⁴Pb than Kerguelen Archipelago lavas (Fig. F10). These Sr, Nd, and Pb isotope data have been interpreted as a result of mixing between the Kerguelen plume and entrained depleted (MORB related) asthenosphere (e.g., Weis et al., 1992).

In contrast, basalts from Site 738 on the southernmost SKP and dredge 8 from eastern Broken Ridge (Figs. F1, F2, F3, F4, F5, F6,) have atypical geochemical characteristics for oceanic lavas. These lavas have very high ⁸⁷Sr/⁸⁶Sr, low ¹⁴³Nd/¹⁴⁴Nd, and very high ²⁰⁸Pb/²⁰⁴Pb and ²⁰⁷Pb/²⁰⁴Pb ratios that accompany relatively low ²⁰⁶Pb/²⁰⁴Pb (Figs. F9, F10). They also have relative depletions in abundances of Nb and Ta (Fig. F11A); moreover, La/Nb correlates positively with ⁸⁷Sr/⁸⁶Sr (Fig. F12). Mahoney et al. (1995) concluded that these isotope characteristics, coupled with depletions of Nb and Ta, arose from a continental lithosphere component that contributed to these basalts, a hypothesis also proposed by Storey et al. (1989) to account for Ta depletion in basalts dredged from the Kerguelen Plateau. Significant relative depletion in Nb is also evident in basalts dredged from the 77° graben on the SKP and from Site 747 on the CKP (Fig. F11A). Trends to anomalously high Th/Nb and La/Nb and a positive correlation between ⁸⁷Sr/⁸⁶Sr and La/Nb are also defined by the Bunbury Basalt, southwest Australia, and the Rajmahal Basalt, northeast India (Figs. F11B, F12). These continental basalts, erupted at ~123-130 Ma and 116 Ma, respectively, are contaminated to varying degrees by continental crust (Frey et al., 1996; Kent et al., 1997). The combination of geochemical features in basalts from Site 738 and eastern Broken Ridge (i.e., very high ⁸⁷Sr/⁸⁶Sr and low ¹⁴³Nd/¹⁴⁴Nd; high ²⁰⁸Pb/²⁰⁴Pb and ²⁰⁷Pb/²⁰⁴Pb that accompany relatively low ²⁰⁶Pb/²⁰⁴Pb; anomalously high Th/Nb and La/Nb; see Figs. F9, F10, F11, F12) is consistent with continental crust as the continental component. In particular, the low ²⁰⁶Pb/²⁰⁴Pb requires aged crust with low U/Pb, such as some types of Archean crust.

In detail, the trend for Site 747 lavas (Fig. F11A) differs from that of other Kerguelen Plateau basalts because Site 747 lavas trend to high La/Nb without elevated Th/Nb. This trend is similar to that for North Atlantic MORB from the lower flow units at Hole 917A (Fig. F11B), which are contaminated by the Archean crust of eastern Greenland, specifically, lower crustal granulite-facies gneiss (Fitton et al., 1998a, 1998b). The effects of continental contamination are very evident in these Hole 917A basalts because their parental magmas were MORB-like with much lower incompatible element abundances than are found in most plume-related lavas. Note that the combination of elevated La/Nb with low Th/Nb is unlike recent estimates of average lower crust composition (e.g., Fig. F11; Rudnick and Fountain, 1995) but is typical of Lewisian granulites (Fig. F11; lower crust estimates are from Weaver and Tarney, 1984). Although not as extreme as some basalts from the lower units of Hole 917A, several geochemical characteristics of Site 747 basalts are consistent with crustal contamination—namely, (1) the trend to high La/Nb without high Th/Nb; (2) the offset to low ¹⁴³Nd/¹⁴⁴Nd from the ⁸⁷Sr/⁸⁶Sr-¹⁴³Nd/¹⁴⁴Nd trend defined by Kerguelen Archipelago lavas; and (3) the low ²⁰⁶Pb/²⁰⁴Pb ratios, which are lower than those of all other lavas from the Kerguelen Plateau, Ninetyeast Ridge, Kerguelen Archipelago, and Heard Island (Figs. F9, F10, F11, F12). These characteristics are consistent with ancient continental crust as the contaminant; the high La/Nb-low Th/Nb (Fig. F11A) trend suggests a component similar to Lewisian granulites. Archean granulites on the conjugate Antarctic and Indian margins (e.g., Black et al., 1992) raise the possibility that fragments of such crust were incorporated into the embryonic Indian Ocean and subsequently sampled during formation of the Kerguelen Plateau. In addition, the Os and Pb isotope ratios of peridotite xenoliths in basalts from the Kerguelen Archipelago are interpreted as reflecting Gondwanan lithospheric mantle that was incorporated into the Indian Ocean mantle during rifting (Hassler and Shimizu, 1998; Mattielli et al., 1999).

The geochemical evidence for a continental component in basalts forming the Kerguelen Plateau and Broken Ridge is consistent with a crustal velocity structure suggesting that the SKP contains a stretched continental fragment (Operto and Charvis, 1995; 1996); this geophysical evidence is at ~58°S in the vicinity of the basalts dredged from the 77° graben and cored at Site 750, whereas Site 738 is much farther south at ~63°S and Site 747 is to the north at ~55°S (Fig. F3, F4). These results suggest that continental lithosphere may be widespread in the Kerguelen Plateau.