SUMMARY

Chronology of Kerguelen Plateau-Broken Ridge Magmatism
Biostratigraphic studies of sediment directly overlying igneous basement at Leg 183 sites provide minimum ages for the volcanic and volcaniclastic rock. Middle Albian (~104.5-106.5 Ma) shallow-water sands and clays overlie inflated pahoehoe flows at Site 1136, suggesting that the age of the lavas is close to the ~110 Ma age of all other basalt samples recovered to date from the SKP. Site 1138 on the CKP yielded undifferentiated Upper Cretaceous claystone and sandstone on top of igneous basement; these sediments are overlain by Cenomanian-Turonian (~93.5 Ma) sandstone. This age is older than, but close to, the ~85 Ma date for basalt at Site 747 (M.F. Coffin et al., unpubl. data). Drilling at the conjugate Broken Ridge region (Site 1141; see Fig. 2) did not provide useful minimum ages because the oldest sediment is Miocene, postdating Eocene separation of Broken Ridge and the CKP and, therefore, much younger than the formation age of Broken Ridge.
The first igneous basement ever recovered from Elan Bank at Site 1137 is overlain by late Campanian (73.5-74.6 Ma) packstone. The lavas and volcaniclastic sediments forming basement are likely to be somewhat older, as the packstone is at the top of a basal sedimentary sequence that thickens markedly to the east of Site 1137. Submarine igneous basement of the NKP was cored for the first time at Site 1139 on Skiff Bank and Site 1140 on the northernmost Kerguelen massif (Fig. 3). On Skiff Bank, chalk at the base of the pelagic sedimentary section is earliest Oligocene (32.8-34.3 Ma) in age. Igneous basement is probably older, as grainstone, packstone, and sandstone lie between it and the overlying pelagic section. Nevertheless, the minimum age is not inconsistent with the oldest rocks from the Kerguelen Archipelago (Giret and Beaux, 1984; K.E. Nicolaysen et al., unpubl. data). At Site 1140, lowermost Oligocene (34.3 Ma) pelagic sediment directly overlies basement, and pelagic sediment of late Eocene age (~35 Ma) is intercalated within basalt flows that form the uppermost basement.

Petrogenesis of Basement Igneous Rocks
Except for Site 1139 on Skiff Bank (NKP), the dominant lava type at all drill sites from diverse domains of the Kerguelen Plateau-Broken Ridge LIP are basalt with transitional to tholeiitic compositions (Fig. 19). These basalts range in inferred age from Albian to early Oligocene and are from the SKP (Sites 738, 749, 750, and 1136), Elan Bank (Site 1137), CKP (747 and 1138), NKP (1140), and Broken Ridge (Sites 1141 and 1142). The basalts have relatively low MgO and Ni contents, and their compositions are not similar to primary melts of peridotite. An example of the important role for fractional crystallization in controlling the compositions of these lavas are basalts from Site 1138, which show a systematic downhole trend to Fe-Ti-rich basalt reaching TiO₂ and total Fe₂O₃ contents of 3.2% and 19.2%, respectively. In general, evolved tholeiitic basalt compositions are typical of many flood basalts. The likely explanation is that the youngest magmas in a LIP, like Kerguelen Plateau-Broken Ridge, must ascend through relatively thick lithosphere, thereby promoting cooling and partial crystallization of the magma. Subsequent segregation of olivine- and pyroxene-rich cumulates then forms the high velocity lower crust that is typical of oceanic plateaus such as the Kerguelen Plateau (Charvis et al., 1993, 1995; Operto and Charvis, 1995, 1996; Farnetani et al., 1996; Könnecke et al., 1998; Charvis and Operto, 1999) and led to the complementary evolved residual melts. In regard to this scenario, it is important to realize that we have only sampled the upper 30-200 m of a ~20-km-thick mafic crust. The basement lavas at Site 1139 (Skiff Bank on the NKP) are an exception to the previous generalization; they form an alkaline lava series ranging from trachybasalt to trachyte and rhyolite; the lowermost flow at Site 1139 is a trachyte (Figs. 34, 35). Similar alkaline lava series have erupted in lower Miocene and Pliocene/Pleistocene time in the Southeast Province of the Kerguelen Archipelago (Weis et al., 1993, 1998; Frey et al., in press). The simplest interpretation is that Skiff Bank, which reaches <500 m water depth, is a submerged island analogous to but slightly older than the Kerguelen Archipelago, also on the NKP but 350 km to the east-northeast (Fig. 3).
An important objective of Leg 183 was to evaluate the role of continental crust in constructing the Kerguelen Plateau-Broken Ridge LIP. Previous evidence pointing to a significant role for continental crust in diverse parts of the LIP includes isotopic and trace element abundance data for basalts from the SKP (Site 738), CKP (Site 747), and basalts dredged from the SKP and eastern Broken Ridge (Figs. 11, 12, 13; Storey et al., 1989, 1992; Mahoney et al., 1995) and the seismic structure of the crust in the northern part of the SKP (Operto and Charvis, 1995, 1996). Some mantle xenoliths in the Kerguelen Archipelago lavas also show evidence for a continental lithosphere component (Hassler and Shimizu, 1998; Mattielli et. al., 1999), as does a trachyte from Heard Island (Barling et al., 1994). Drilling at Site 1137, however, recovered the strongest evidence to date for a component of continental crust in the Kerguelen Plateau. A ~26-m-thick fluvial conglomerate (Fig. 44) intercalated with basaltic basement contains clasts of trachyte, rhyolite, granitoid, and garnet biotite gneiss (Fig. 45); the garnet-biotite gneiss, in particular, indicates that continental crustal rocks were once exposed at Elan Bank. Furthermore, although it is difficult to use shipboard geochemical data to identify continental material in mantle-derived basaltic rocks, our study of Site 1137 cores builds a compelling case. The basement basalts at Site 1137 are geochemically distinctive; they have atypically high Zr/Y and Zr/Ti (Fig. 24) and a slight relative depletion in Nb abundance (Figs. 24, 25)--both characteristics are consistent with a component derived from continental crust.
A Nb/Y vs. Zr/Y plot has been used to distinguish between lavas derived from the Icelandic plume and North Atlantic MORB (Fitton et al., 1997, 1998). In this plot (Fig. 31), it is important to realize that different plumes are geochemically distinct; in particular, lavas of varying age, from ~82 Ma to Pliocene-Pleistocene, associated with the Kerguelen plume define a Nb/Y-Zr/Y trajectory along the lower boundary for the Icelandic plume (Fig. 31). We conclude that the Kerguelen plume plots on this line in the lower left portion of the figure. Basement basalts from several locations on the Kerguelen Plateau (and Broken Ridge?) also lie along this trend (namely, Sites 747, 1138, 1138, and 1140). Two other locations are in the Icelandic field (Sites 749 and 750), and two others lie in the MORB field (Sites 738 and 1137); however, basalts from these latter two sites do not have MORB geochemical characteristics (Figs. 9, 25). We infer that basalts from Sites 738 and 1137 are in the MORB field because they are plume-derived basalts that have been contaminated by continental crust (Fig. 25). The effects of crustal contamination are obvious in the Sr, Nd, and Pb isotopic characteristics of Site 738 basalts (Figs. 9, 10). Isotopic data are not yet available for Site 1137 basalts, but the clasts of garnet-biotite- gneiss in a conglomerate intercalated with these basalts (Fig. 45) unambiguously show that continental crust is present in the oceanic environment of Elan Bank.
Continental material, whether derived from continental crust or subcrustal continental lithosphere, is occasionally incorporated into oceanic lithosphere. At one end of the spectrum are microcontinents such as Seychelles and Jan Mayen, which maintain normal continental crustal thicknesses when isolated in ocean basins by jumps of seafloor spreading centers, with or without involvement of mantle plumes. At the other end of the spectrum are subtle geochemical signatures in otherwise typical mid-ocean ridge basalts that have been interpreted to represent the influence of deeply recycled continental material or subcrustal continental lithosphere (e.g., Mahoney et al., 1996). As described above, interpreted continental components of the Kerguelen Plateau and Broken Ridge encompass a range of petrologic, geochemical, and geophysical signatures and, therefore, may be accounted for by more than one process. However, the simplest mechanism for incorporating continental material into the Kerguelen Plateau-Broken Ridge LIP is a ridge jump postdating the breakup and initial seafloor spreading between India and Antarctica. One or more ridge jumps to the north would have transferred continental parts of the Indian plate to oceanic portions of the Antarctic plate, accounting for features like Elan Bank and the portion of the SKP characterized by continental crustal velocities. Published plate motion models (Fig. 2) do not show any ridge jumps because the oceanic crust between Antarctic and Kerguelen Plateau formed during the long Cretaceous Normal Superchron and, therefore, is not datable using the usual technique of marine magnetic anomaly identification. Nevertheless, the unambiguously continental garnet-biotite gneiss recovered as clasts in a conglomerate intercalated with basement basalt at Site 1137 on Elan Bank strongly suggests that at least one northward ridge jump transferred a continental fragment (Elan Bank) from the Indian to the Antarctic plate.

Environmental Impact of Kerguelen LIP Volcanism
Evidence from basalts and overlying sediments at Sites 738, 747, 748, 749, and 750 (Fig. 1), combined with results of subsidence modeling (Coffin, 1992), shows that much of the igneous crust of the SKP and CKP was erupted in a subaerial environment. Portions of the SKP remained subaerial for as much as 50 m.y. after volcanism ceased. Leg 183 drilling results corroborate, extend, and add detail to those previous results. On the SKP (Site 1136), we cored upper bathyal to neritic sediment overlying inflated pahoehoe lavas up to 20 m thick. The basalts lack features of submarine volcanism (e.g., pillows and quenched glassy margins) suggesting subaerial eruption. The CKP (Site 1138) was above sea level during the final stages of construction; subaerial pyroclastic flow deposits overlie ~5-m-thick subaerial lava flows ranging from inflated pahoehoe to classic aa. Terrestrial and shallow marine sediment containing wood fragments, a seed, spores, and pollen overlies igneous basement, documenting for the first time that the CKP was subaerial after volcanism ceased. In a conjugate position to the CKP, Broken Ridge basaltic lavas were erupted in subaerial (Site 1141) and possibly submarine (Site 1142) environments.
The igneous basement complex of Elan Bank at Site 1137 consists of seven basaltic lava flows and three sedimentary units. Six of the seven 7- to 27-m-thick lava flows were erupted subaerially, as indicated by oxidation zones and inflated pahoehoe characteristics, and the other may have erupted into wet sediment. Some of the interbedded volcaniclastic sedimentary rocks were deposited in a fluvial environment (braided river), consistent with subaerial eruption of the basalts. Neritic packstones overlying the igneous basement complex, in turn succeeded by pelagic oozes, indicate gradual subsidence of Elan Bank.
Skiff Bank (Site 1139) was also subaerial during its final stages of formation, as indicated by a succession of volcanic and volcaniclastic rocks (some of which are oxidized) underlain by fourteen 2- to 20-m-thick lava flows, including both pahoehoe and aa types. After volcanism ceased, paleoenvironments of the overlying sediments changed intertidal (beach deposits) to very high energy, near-shore (grainstone and sandstone) to low-energy, offshore (packstone) to bathyal pelagic (ooze). In contrast, igneous basement Site 1140 at the northernmost tip of the NKP consists entirely of pillow basalts and intercalated pelagic sediment. However, seafloor depths at all six other basement sites drilled during Leg 183 are between 1000 and 2000 m, whereas Site 1140 is situated at a water depth of 2450 m.
An unexpected result of Leg 183 drilling was the discovery that highly evolved, felsic magmas were erupted explosively during the final stages of magmatism over extensive regions of the Kerguelen Plateau. Igneous basement recovered from four previous ODP sites on the SKP and CKP did not include felsic lavas, but at four Leg 183 drill sites we recovered pyroclastic flow deposits and dense lava samples of trachyte, dacite, and quartz-bearing peralkaline rhyolite. Alkalic lavas have not been previously recovered from the basement of the Kerguelen Plateau. At Site 748 on the SKP, however, alkalic basalt was recovered from ~200 m above basement. At Site 1137 on Elan Bank, a 15-m-thick sanidine-rich vitric tuff is intercalated between basaltic lava flows. Well-preserved bubble wall glass shards in part of the tuff together with abundant broken crystals indicate that the tuff formed in an explosive volcanic eruption. Higher in the stratigraphic sequence at Elan Bank, a fluvial conglomerate contains clasts of rhyolitic and trachytic lavas. At Site 1138 on the CKP, we recovered a 20-m-thick volcaniclastic succession containing six trachytic pumice lithic breccias that were deposited by pyroclastic flows. This volcaniclastic sequence also includes highly altered ash fall deposits that contain accretionary lapilli. Above this sequence we recovered a reworked deposit of rounded cobbles of flow-banded dacite. At Site 1139 on Skiff Bank, which forms part of the NKP, the uppermost basement contains a variety of felsic volcanic and volcaniclastic rocks. In contrast to the CKP and Elan Bank sites, biostratigraphic ages of sediments directly overlying this basement suggest that this episode of felsic volcanism is Cenozoic in age. The Skiff Bank section includes densely welded pyroclastic flow deposits of quartz-bearing peralkaline rhyolite, in addition to lava flows and reworked cobbles of volcanic rock ranging from sanidine-rich trachyte to rhyolite.
Evolved magmas (e.g., trachyte, phonolite, and peralkaline rhyolite) are erupted during plume-related volcanism at oceanic islands and in some continental flood basalt provinces (e.g., Parana, Etendeka, Karoo, and Siberian Traps). Typically these eruptions occur near the end of voluminous, basaltic magmatism. Two alternative modes of formation for highly evolved magmas are partial melting of lower crustal rocks or as residual magmas created as the supply of mantle derived basaltic magma wanes, leading to formation of crustal-level magma chambers in which highly evolved magma forms through crystal fractionation (±wallrock assimilation) of basalt.
The eruption of enormous volumes of basaltic magma during formation of the Kerguelen Plateau-Broken Ridge LIP probably had significant environmental consequences because of the release of volatiles such as CO₂, S, HCl, and HF. A key factor in the magnitude of volatile release is whether the eruptions were subaerial or submarine; hydrostatic pressure inhibits vesiculation and degassing of magma during submarine eruptions. Results of Leg 183 drilling complement earlier results from Legs 119 and 120 in demonstrating that subaerial basaltic eruptions occurred during the final constructional stages of the plateau.
Another important factor that would have increased the environmental consequences of Kerguelen Plateau-Broken Ridge LIP volcanism is the high latitude at which the plateau formed. In most effusive basaltic eruptions, released volatiles remain in the troposphere. However, at high latitudes, the tropopause is relatively low, allowing large mass flux and basaltic fissure eruption plumes to transport SO₂ and other volatiles into the stratosphere (e.g., 1783-1784 Laki eruption in Iceland). The sulfuric acid aerosol particles that form in the stratosphere after such eruptions have a longer residence time and greater global dispersal than if the SO₂ remains in the troposphere; therefore, they have greater effects on climate and atmospheric chemistry. The large volume and long duration of subaerial basaltic volcanism on the Kerguelen Plateau-Broken Ridge LIP, combined with the high latitude of most of the plateau, would all have contributed to potential environmental effects.
Highly explosive felsic eruptions, such as those that formed the pyroclastic deposits on Elan Bank, Skiff Bank, and the CKP can also inject both particulate material and volatiles (SO₂, CO₂, possibly HCl) directly into the stratosphere. The previously unrecognized, significant volume of explosive felsic volcanism that occurred when the Kerguelen Plateau and Broken Ridge were subaerial would have further contributed to the effects of this LIP volcanism on global climate and environment. The total volume of felsic volcanic rocks is still poorly constrained, but our results indicate that they account for a significant fraction of the volcanic deposits erupted during the final stages of magmatism at several locations on the Kerguelen Plateau.

Tectonic History
The uppermost volcanic basement of the Kerguelen Plateau was mostly erupted in a subaerial environment (Leg 183 Sites 1136, 1137, 1138, 1139, and 1141; Leg 119 Site 738; Leg 120 Sites 747, 749, and 750), although bathymetrically deeper basalts on the NKP (Site 1140) formed under water. Sediments overlying basaltic basement at Leg 183 drill sites record the vertical tectonic history of the Kerguelen Plateau through changing facies. At Site 1136 on the SKP, neritic clay and sand overlie basement, which are in turn overlain by pelagic chalk and ooze. This sedimentary succession documents subsidence of the SKP since Early Cretaceous time. On the CKP (Site 1138), terrestrial and shallow marine sediment overlie basement. In particular, the transition from oxidized neritic sediment to black claystone to pelagic sediment likely records both thermal subsidence and eustatic sea level rise. At Site 1137 on Elan Bank, fluvial sediment is interbedded with basalt, and neritic packstone succeeded by pelagic ooze overlies volcanic basement. Despite a component of continental crust in Elan Bank, its subsidence behavior appears to resemble that of oceanic crust. The subsidence of Skiff Bank (Site 1139), part of the NKP, is recorded by a transition from subaerial (volcanic and volcaniclastic rocks) to intertidal (beach deposits) to very high-energy, near-shore (grainstone and sandstone) to low-energy offshore (packstone) to bathyal pelagic (ooze). In summary, during Leg 183 we recovered sediments that record subsidence of the SKP, CKP, Elan Bank, and Skiff Bank; for the latter three sectors, these constitute the first such records.

Postcruise Research
The basement rocks will be the focus of several shore-based studies; for example:

1. The chronology of basement rocks will be principally determined by ⁴⁰Ar/³⁹Ar studies of relatively unaltered whole rock samples and potassium-bearing phases such as plagioclase and sanidine.
2. The composition (major and trace element abundances) and isotopic ratios (Sr, Nd, and perhaps others) of unaltered phenocrysts (such as plagioclase and clinopyroxene) will provide information on parental magma composition and the role of crustal processes such as fractional crystallization, magma mixing, and assimilation of wallrocks.
3. The composition (major and trace element abundances) and isotopic ratios of (O, Sr, and perhaps others) in secondary phases formed as the lavas interacted with the surficial environment will be used to understand the postmagmatic processes in subaerial and submarine environments.
4. The composition (major and trace element abundances) and isotopic ratios (O, Sr, Nd, Pb, Hf, and Os) of whole rocks will be used to understand both magmatic and postmagmatic processes and the role of geochemically distinct mantle and crustal components in these rocks.
5. The magnetic inclination of basement rocks will be determined, providing paleolatitude information necessary for plate kinematic and paleoceanographic studies.
6. The range of lava flow morphology and eruption style will be used to understand the physical volcanological processes that created the Kerguelen Plateau-Broken Ridge LIP.
7. Shore-based research will also focus on unique core samples that bear particularly on narrowly focused problems; for example:
a. Both the chemical and physical characteristics of the felsic volcanic units will be used to understand the environmental effects of such volcanism, especially the units that reflect explosive eruption of presumably volatile-rich magmas.
b. The diverse array of clasts (basalt, rhyolite, trachyte, granitoid, and garnet biotite gneiss) in the 26 m of conglomerate within the basement at Site 1137 will be used to define the depositional environment and geologic terrain, in terms of geochemistry and age, existing at Elan Bank near the end of its volcanic history; this research will include U-Pb dating of zircons in the gneiss clasts and volcaniclastic units.
c. The glassy pillow rinds at Site 1140 will be studied to determine the composition of the basaltic melts; such samples are particularly important for determining the volatile contents of the magmas at this NKP location.
8. Tectonic studies will incorporate new age dates of basement rocks and the existence of continental material in developing new, improved regional plate reconstructions.
9. Seismic volcanostratigraphy will extend borehole results from basement drilling regionally to interpret different types of volcanism and relationships between magmatism and tectonism;
10. Downhole logging data, the Formation MicroScanner (FMS) in particular, will be used to determine true lava flow thicknesses and structure in intervals of low core recovery;
11. Physical properties and downhole logging data will be used to better understand the effects of intrabasement velocity inversions on seismic reflection data.

1. The Turonian-Coniacian black shale at Site 1138 will be the focus of several studies including analyses for a wide range of metals to evaluate abundance anomalies of metals that may have been introduced into seawater by hydrothermal activity associated with LIP formation, C and N isotopic analyses of the organic carbon fraction, which can be used as proxies for source and productivity followed by biomarker molecule extractions and analyses.

2. The expanded sections (e.g., middle Eocene to uppermost Paleocene at Site 1135 and lower to middle Eocene at Site 1136) will be used to understand high latitude paleoceanographic conditions and improve biostratigraphic resolution.

3. Paleomagnetic studies will be used to establish magnetostratigraphy in sedimentary sections.

4. Palynological studies will help date terrestrial and terrigenous sediment and determine vegetation types and paleoclimates when the Kerguelen Plateau and Broken Ridge were subaerial.

5. Exploratory research on isotopic variations of Li and first-series transition elements in the sedimentary environment.

6. Seismic stratigraphy will use borehole results to interpret key seismic reflections regionally, enabling dating of tectonic and paleoceanographic events.

7. Subsidence analyses will model the Kerguelen Plateau's vertical tectonic history, taking different magmatic and deformational events into account.

Leg 183 Drilling Summary (Table 1)

Leg 183 Figures

Leg 183 Operations Synopsis
Leg 183 Table of Contents