ABSTRACT
The circulation of cold, deep Antarctic Bottom Water (AABW) is one of the controlling factors in Earth's climate. Today, 40% of this water enters the world ocean through the Southwest Pacific Gateway as a thermohaline driven Deep Western Boundary Current (DWBC). South of 46°S, the DWBC is coupled with the wind-driven Antarctic Circumpolar Current (ACC). Understanding the evolution of the Pacific DWBC is fundamental to understanding world oceanic and climatic histories. The ACC-DWBC system has evolved since the late Oligocene (32-20 Ma), when plate movements created the first deep-water oceanic gaps south of Australia and South America. An excellent stratigraphic record of these events, and of the development of the modern ACC-DWBC, occurs in sediment drifts east of the New Zealand microcontinental plateau. Seven Southwest Pacific drill sites are proposed to reconstruct the stratigraphy, paleohydrography, and dynamics of the DWBC and related water masses. The proposed sites comprise a transect of water depths from 320 m to 4460 m, and span a latitudinal range from 39°S to 51°S. Only one previous hydraulic piston core site is located in this large region (Deep Sea Drilling Project [DSDP] Site 594), and earlier DSDP drilling (Sites 275-276) occurred at sites where Neogene sediment is largely missing due to DWBC erosion. Consequently, our knowledge of Southwest Pacific ocean history, and of the development of the ACC-DWBC system, is extremely poor. Coring during Leg 181 will provide the sedimentary sequences needed to study a range of high-priority problems in Southern Ocean Neogene paleohydrography, sedimentology, paleoclimatology, and micropaleontology.
1.The Pacific Deep Western Boundary Current (DWBC) is the largest single contributor today to the deep waters of the world's oceans (20 Sverdrups [1 Sv = 106 m3s-1]; Fig. 1), and deciphering its history is, therefore, of fundamental importance to global ocean hydrography.
2.The stratigraphic record of the eastern New Zealand Plateau and its abyssal margins is the best one available for deciphering the development history of Pacific Southern Ocean water masses and the sediment drifts they deposited. Recent publications (Carter and Carter, 1993; Lewis, 1994; Carter and McCave, 1994; L. Carter et al., 1996) delineate the region between the Solander Trough and the Kermadec Trench, east of the modern Australian-Pacific plate boundary, as an integrated sediment source-transport-sink area, termed the Eastern New Zealand Oceanic Sedimentary System (ENZOSS). Since ~10 Ma, sediment from mountains along the New Zealand plate boundary has been transported through deep-sea channel/fan systems, delivered into the path of the DWBC, entrained northwards within this current system, and finally consumed by subduction at the same plate boundary after a transport path of some 3500 km. The stratigraphic record from the ENZOSS and, in particular, any new high-resolution Neogene Ocean Drilling Program (ODP) sections from its deep-sea parts, are directly relevant to one of the most important unresolved problems of Cenozoic climatology‹namely, the timing and precise nature of the development of widespread glaciation on the Antarctic continent (e.g., Barrett, in press). In turn, the same glacial events that contribute source water to the DWBC and its companion flow, the Antarctic Circumpolar Current (AAC), force the boundary current south of 49°S.
3.The gateway region includes two major oceanic fronts, the Subtropical Convergence and the Subantarctic Front, and is proximal to a third, the Antarctic Convergence (Figs. 2, 3). Thus, the region is in a prime position to allow determination of the migration of these boundaries, the forcing processes which cause them to move, and the environmental response to their movement.
Modern Oceanography
The supply of deep water to the Pacific Ocean is dominated by a single source‹the DWBC, which flows north out of the Southern Ocean along the east side of the Campbell Plateau-Chatham Rise-Hikurangi Plateau located east of New Zealand (Figs. 1, 2, 3, 4, 5, and Table 1). The volume transport of the DWBC is about 20 Sv, which comprises ~40% of the total input of deep water to the world's oceans (Warren, 1973; 1981). (A secondary, but, minor, flow of ~3 Sv of deep-water flows north into the Peru-Chile basin; Lonsdale, 1976). The magnitude of DWBC flow, and the low temperature of the water involved, are major determinants of the oceanography of the Pacific Ocean and of the global heat balance. Monitoring the DWBC flow at its entry into the Pacific is a key area where the "global salt conveyor" hypothesis (Gordon, 1988; Broecker et al., 1990; Schmitz, 1995) can be tested, as the flow, thereafter, is believed to spread out to fill the Pacific. Some water upwells and returns at shallower depths to the Atlantic, whereas other waters return south as North Pacific Deep Water (NPDW).
The supply of cold water to the deep Pacific from the main generating regions in the Weddell and Ross Seas is modulated by the ACC, which mixes these waters with North Atlantic Deep Water (NADW) in the South Atlantic to form Circumpolar Deep Water (CDW). Deep-water output to the Pacific, therefore, carries the combined signatures of Southern Ocean processes in the region of deep-water formation, chemical composition related to Southern Ocean gas exchange, and NADW. Despite its turbulent passage around Antarctica, CDW is not completely mixed, and a distinct NADW salinity-maximum can be recognized at depths of 2800 m (at 55°S) deepening northwards to 3400 m (at 28°S). In the southwest Pacific, the DWBC comprises three main divisions (1) lower CDW, a mixture of bottom waters generated around Antarctica, in particular cold Weddell Sea deep water and NADW; (2) salinity-maximum middle CDW, representing the NADW core; and (3) strongly nutrient-enriched and oxygen-depleted upper CDW, mainly derived from Indian Ocean outflow added to Pacific outflow returning through the Drake Passage. The DWBC has its upper boundary at depths around 2000-2500 m. On the eastern side, the DWBC is overlain between 2550 and 1450 m depth by south-flowing NPDW, and is marked by high silica and an oxygen minimum. Regionally, both DWBC and NPDW are overlain by low-salinity Antarctic Intermediate Water (AAIW) (Fig. 4 and Table 1).
The ACC-DWBC enters the Southwest Pacific through gaps in the Macquarie Ridge complex before passing along the 3500-m-high margin of the Campbell Plateau. Near the mouth of the Bounty Trough, the ACC uncouples and continues its eastward path, whereas the DWBC flows north around the eastern end of Chatham Rise and through Valerie Passage, where a small part of the flow diverges through gaps in the Louisville Ridge. Valerie Passage, the 250-km-wide gap between the Chatham Rise and the Louisville seamount chain, therefore marks the gateway to the Pacific for the DWBC.
Sedimentary Record of the ACC-DWBC
Sediments on the eastern New Zealand margin at shelf to upper bathyal depths (100-1000 m) are known to have been strongly affected by currents since at least the late Oligocene (Ward and Lewis, 1975; Carter, 1985; Fulthorpe et al., 1996). This evidence for strong paleoflows, together with the confirmation that substantial Antarctic glaciation commenced at least as early as the early Oligocene (Shackelton and Kennett, 1975; Barrett, in press; Prydz Bay drilling, Barren and Larsen, 1989), implies that Pacific hydrography has been fundamentally affected by an evolving circumpolar current and western boundary current system since the mid-Cenozoic.
To reconstruct the paleoflow of the DWBC and overlying current system requires drill sites through thick undisturbed fine-grained sediment masses constructed under the influence of the current. Seismic records indicate the presence of suitable sedimentary drifts at many points along the eastern edge of the New Zealand Plateau in water and paleowater depths between 300 m and 5500 m (Carter and McCave, 1994; L. Carter et al., 1996). Three sediment sources are involved in building these drifts: (1) transport into the area via the DWBC itself (e.g., subantarctic diatoms present in the drifts at 40°S; Carter and Mitchell, 1987); (2) pelagic and hemipelagic rain and airfall rhyolitic ash which, over the last 20 k.y., has been input at a rate of up to one third that of fluvial terrigenous sediment (Carter et al., 1995); and (3) terrigenous sediment from New Zealand, deposited at the shelf-edge by mid-accretion (Fulthorpe and Carter, 1991) or delivered into the path of the DWBC from turbidity currents traveling down the Solander, Bounty, and Hikurangi channel systems. Each of these sediment sources can be constrained, and the sedimentary dynamics and transport paths of the modern system are moderately well delineated (e.g., Carter and Carter, 1993; Carter and McCave, 1994; Lewis, 1994). In contrast, little is known regarding the earlier Cenozoic record of the DWBC.
The available seismic records show that the DWBC has been active along the eastern New Zealand margin since at least the Miocene, and probably since the mid-Oligocene (32 Ma) (Carter and McCave, 1994). After ~10 Ma, abundant terrigenous material was shed from rising mountains along the Alpine Fault plate boundary (Kennett, von der Borch, et al., 1986) and fed into the Solander, Bounty, and Hikurangi channel systems, especially at times of late Neogene glacial sea-level lowstand. Much of this sediment was then entrained into the DWBC drift system, which carried it northward to be eventually subducted into the Kermadec Trench.
Sediment is delivered into the DWBC through two newly described transport conduits, the Bounty (Carter and Carter, 1993) and Hikurangi (Lewis, 1994) channel-fan systems. A third feeder channel, Solander, is poorly known, but extends for >450 km before discharging into the DWBC at Emerald Basin between Macquarie Ridge and the Western side of Campbell Plateau (R. Carter et al., 1996). The Hikurangi Fan has been termed a "fan-drift" by Carter and McCave (1994) because it apparently represents the extreme case of a fan whose thickness and facies pattern are directly remolded by a deep current into the form of a sediment drift. In contrast, the Bounty Fan, located in a bathymetric embayment, has retained its fan morphology and has developed directly across the path of the DWBC (Carter and Carter, 1993), the only evidence of drift formation being scour of the northern fan and redeposition of material as a series of small discrete ridges. Compared to Hikurangi Fan Drift, the Bounty Fan has formed in a region where the DWBC is slowed because of a gently sloping western boundary and the shelter provided by Bollons Seamount.
The two described abyssal fans are supplied with sediment by turbidites passing through the Bounty and Hikurangi channels, each of which is over 1000 km long. The Hikurangi channel heads in the Kaikoura Canyon, only a few hundred meters from shore, and less than ten kilometers from the rapidly rising 2.5-km-high Seaward Kaikoura Mountains. The Hikurangi system is therefore active today during interglacial times. In contrast, the Solander and Bounty Channels head in a number of canyons that indent the edge of the continental shelf. The Bounty and Solander systems are, therefore, strongly sea-level (i.e., climatically) controlled with most sediment being fed into them during glacial lowstands. In contrast, during interglacial periods the same sediment stream is diverted along the inner shelf, some of it even reaching the Hikurangi System via the Kaikoura Canyon (Carter and Herzer, 1979).
1. Delineate the Cenozoic development of zonal water masses and the ACC system. Current understanding of paleoclimate suggests that the earliest major meltwater events from Antarctic glaciation occurred at ~38 Ma (Eocene/Oligocene boundary; Shackleton and Kennett, 1975; Miller et al., 1990). At 32 Ma (mid-Oligocene), the South Tasman Gateway opened, including the Balleny Fracture zone (Lonsdale, 1988), thereby connecting the Indian and Pacific Oceans for the first time (Kennett, et al., 1972). Finally, it was not until ~20 Ma (earliest Miocene) that the opening of Drake Passage (Boltovskoy, 1980) allowed the establishment of the full pattern of circum-Antarctic ocean circulation. The evolution of this system, including periods when the boundary current component may have extended into shallow depths or reversed (Mikolajewicz et al., 1993), is the target of Sites SWPAC-5A and 6B. These sites may also penetrate to the regionally widespread 29 Ma mid-Oligocene unconformity (Marshall Paraconformity; Carter, 1985), the genesis of which may relate to the inception of ACC-DWBC activity as much as to global sea-level change (cf. the postulated large 29 Ma lowstand of Haq et al., 1987). Site SWPAC-1A is targeted on large Miocene-Pliocene platform drifts that grew from a paleowater depth of 1000 m in the head of the Bounty Trough (Fulthorpe and Carter, 1991).
2. Infer the changing paleohydrography of the CDW supply to the Pacific Ocean; in particular, to trace the history of mixing of Weddell Sea Deep Water (WSDW) and NADW components.
Measurement of delta18O, delta13C and Cd/Ca from microfossil tests will be used to distinguish the relative contribution of nutrient-enriched NADW and 13C depleted components
(Boyle and Keigwin, 1987; Oppo and Fairbanks, 1987; Charles and Fairbanks, 1992; Bertram et al., 1995). Sites SWPAC-2A and 5A were selected to maximize the chance of obtaining the high-quality carbonate-rich
records required for such measurements. For the deeper water sites, should carbonate percentages be low, the bulk carbonate technique of Shackleton et al. (1993) may yield a satisfactory isotope stratigraphy.
However, we have obtained monospecific benthic and planktonic oxygen isotope records for the mid-Holocene to Stage 3 from a core at 4802 m water depth, and McCave and Carter (in press) estimate the carbonate compensation depth (CCD) to lie at ~4750 m. The late Neogene stratigraphy may be strongly supported by a tephrochronology derived from the numerous widespread Cenozoic ash beds deposited east of New Zealand. Global understanding of the history of these water masses will require the comparison of gradients of water composition between sites in the North Atlantic, South Atlantic, North Pacific, and Southern Oceans.
3. Determine the relative paleoflow speeds of deep and intermediate waters, and thereby estimate the changing flux of CDW into the Pacific through time.
The noncohesive sortable silt (10-63 micron) fraction has been shown at widely separated locations to yield coherent indications of flow speed and, hence, water mass movement (McCave et al., 1995a, b; Manighetti and McCave, 1995; Robinson and McCave, 1994; Haskell et al., 1991). Evaluation of such grain-size signals in the North Chatham Drift (Site SWPAC-5A) and Campbell Drift (Site SWPAC-7A) will permit estimation of the velocity behavior of the DWBC. Sites 1C, 2A, and 6B will yield indications of the behavior of low salinity AAIW. Site SWPAC-7A, located a little to the south of the Bounty Trough, will be important for assessing the extent to which the ACC acts as a driving force for CDW inflow, because it is at the latitude of the mouth of the trough that the modern ACC veers east into the Pacific where it decouples from the deep boundary flow (cf. Semtner and Chervin, 1992).
4. Establish the history and depth ranges of AAIW across the New Zealand Plateau.
In the North Atlantic, an intermediate water (possibly paleo-Labrador sea water) has been shown to increase both in depth range and speed during glaciation, concomitant with a decrease in NADW production (Boyle and Keigwin, 1987; Manighetti and McCave, 1995). Since this change is associated with suppression of the NADW, there is little reason to expect the same glacial/interglacial changes to AAIW in the southern ocean. However, Pudsey et al. (1988) have argued that AABW production also diminished during glacial periods as a result of the grounding of ice sheets, in which case the thickness of AAIW may well have increased concomitantly. If the vigor of global deep circulation was decreased by these North Atlantic and Antarctic events, then, during glacial times, the Indian/Pacific upper CDW should have become even more nutrient enriched and oxygen depleted than it is today. Material from Site SWPAC-5A (water depth 3308 m) will be used for delta13C and trace element analysis (e.g., Cd/Ca in calcite and opal) to allow ocean paleochemistry to be used to determine whether during glaciation periods the site lay under severely depleted AAIW or enriched CDW.
5. Determine the history of productivity and surface water mass fluctuations in the vicinity of the Subtropical Convergence.
Near zones of upwelling, such as the subtropical convergence (STC), it is usually difficult to distinguish between climatically controlled temporal and spatial changes in productivity because the convergence moves. However, for at least the last full glacial/interglacial cycle the STC has apparently been topographically trapped over the Chatham Rise (Fenner et al., 1992; Nelson et al., 1993). This raises the prospect of being able to obtain temporal records of productivity change from faunal, isotopic, and chemical data without the aliassing usually produced by shifts in the position of such convergences. ODP Site 594, in 1200 m of water just south of the present STC is a valuable control because it shows that cold water reached there during the last glaciation (Nelson et al., 1993), probably representing waters wind-drifted from the Subantarctic Front (SAF), which itself remained bounded by the Campbell Plateau. Site 594 provides a high-quality record extending to the middle Miocene, and we aim to match it by similar records from beneath the STC and farther north (Site SWPAC-5A) and south (Sites SWPAC-6A, 7A).
6. Examine the shifting positions of the Subantarctic Convergence.
The zone of cold water between the STC and the Antarctic Convergence (AAC) at 60°S is divided by the SAF at about 51°S (Health, 1995). Nelson et al. (1993) found a sharp cooling of waters at glacial levels in Site 594, suggesting that the SAF lay nearby. Our most southerly drift sites (Sites SWPAC-6B, 7A) and Site SWPAC-8A on the levee of the Bounty Channel will allow us to assess the shift in position of these climatically important fronts, using faunal and coarse-fraction analysis, and stable isotopic and magnetic susceptibility measurements to trace ice-rafted detritus.
7. Test the record of circum-Antarctic flow against the Milankovitch orbital model, including estimates of simultaneity with Northern Hemisphere records.
Achieving this objective requires the retrieval of high-quality, long-term faunal and isotopic records from Southern Ocean sites to assess changes in temperature, salinity, and CO2 as components of the climate system. The relative timing of events between the Northern and Southern Hemispheres will be evaluated (cf. Nelson et al., 1985). We anticipate the best long-term records will come from Sites SWPAC-2A and 5A on the north flank of Chatham Rise, where the North Chatham Drift is up to 1000 m thick and probably extends back to the late Oligocene. The deeper levels of the drift will be principally examined farther north (Site SWPAC-9A), where the thickness is reduced, although obviously with less stratigraphic resolution.
8. Study the effect of an oscillating sediment source controlled by Pleistocene sea-level cyclicity on fan overbank turbidite deposition and sediment supply to the DWBC.
The Bounty Channel and Fan are fed with terrigenous sediment through a number of submarine canyons, which cut the eastern South Island shelf edge. Located about 30-60 km offshore, these canyons were alternately supplied directly with sediment during the Pliocene-Pleistocene glacial lowstands, and cut off during interglacial highstands when sediment was moved directly along the inner shelf (Carter and Carter, 1993). Consequently, the Bounty Fan was supplied with terrigenous sediment mostly during glacial periods, and the levees of the Bounty Channel comprise a regular sequence of 5 to 8-m-thick packets of glacial silt-mud turbidites alternating with interglacial biopelagic calcareous ooze (Carter et al., 1990). The regularity of these cycles is such that they can be matched prima facie with the oxygen isotopic record back to Stage 100, and viewed as the deep-sea record of continental shelf stratigraphic sequences (Carter and Carter, 1992). Site SWPAC-8A will penetrate about 50 of the cycles identified on seismic records, providing (1) a high-quality record of turbidity current activity through the Bounty Channel since the early Pleistocene; (2) a test of the correlation between the observed seismic cyclothems and the oxygen isotopic stages; and (3) a quantitative model of rates of sediment supply to the DWBC and deep sea during a time of glacio-eustatic oscillation of sea-level.
9. Estimate the quantities of sediment fed into the DWBC from all sources, including terrigenous sediment delivered through the Solander, Bounty, and Hikurangi channels to understand the relative importance of tectonic, climatic, sea-level and water-mass controls and to derive a sedimentary budget for the ENZOSS.
Since the Pliocene the three largest sources of sediment for the DWBC drifts have been turbidity currents traveling the length of the Solander, Bounty, and Hikurangi channel systems (Carter et al., 1990; 1994; and 1996; Lewis, 1994). Other major sediment sources are direct transport into the region by the DWBC and pelagic, hemipelagic, and volcanic fallout. In addition to the differing primary targets at each site, all SWPAC sites will contribute data toward the development of a quantitative sedimentary model and budget for the ENZOSS.
Some of the drift sequences targeted for drilling, particularly those nearer North Island, will contain Pliocene-Pleistocene volcanic ash layers of potential value for correlation. Several very large explosive eruptions have occurred in the Central Volcanic Region of New Zealand in the last few million years (Shane et al., 1996). The largest eruptions in the last 50 k.y. have exceeded 100 km2 in volume. These ashes are generally distributed to the east of New Zealand (Stewart and Neall, 1984), and occur widely in marine cores (Ninkovitch, 1968; Lewis and Kohn, 1973; Watkins and Huang, 1977; Kyle and Seward, 1984; Nelson, 1988; Carter et al., 1995). Volcanic ash horizons may also provide useful stratigraphic markers at horizons as old as late Miocene (van der Lingen, 1968) and perhaps earlier.
The application of combined isothermal plateau fission-track (ITPFT) for dating and laser ablation inductively-coupled plasma mass spectrometry (LA-ICP-MS) for fingerprinting has revolutionized the ash chronostratigraphy of the Pliocene-Pleistocene Wanganui Basin (e.g., Naish et al., 1996), and holds great promise for application to any ash layers that are recovered from Leg 181 sites.
DRILLING STRATEGY AND PROPOSED SITES
1. Chatham Rise Transect (A)
This transect comprises three holes in water depths between 300 and 3308 m located in a northeast band across major sediment drifts of the Chatham Rise. All sites are above the modern CCD, recently determined to be at ~4750 m in this area (McCave and Carter, in press). Sections span an inferred age of early Miocene to Holocene within a thickness of 400-750 m. Two sites (Sites SWPAC-2A and 4A) terminate in presumed Oligocene sediment, which marks the regional inception of strong bottom-water flow into the Pacific.
2. Campbell Plateau Transect (B)
This transect commences with Site SWPAC-1A in 300 m of water on the eastern South Island shelf (Canterbury Drifts; to sample upper AAIW and basal thermocline water), then passes southeast of DSDP Site 594 to link to two sites near the eastern edge of the Campbell Plateau, Sites SWPAC-6B and 7A. Site SWPAC-6A (960 m water depth) is targeted to sample AAIW. Biopelagite is present at the surface and is expected to extend downhole to the target depth. Sedimentation rate in a nearby core (F-121) is <1 cm/yr, but, from the thickening of seismic intervals toward Site SWPAC-6A, significantly higher rates are expected there. Site SWPAC-7A is situated on the crest of the large Campbell Drift, and at 4390 m lies close to the regional CCD. However, a benthic and planktonic isotope record from a site to the north at 4802 m, and pilot studies of bulk carbonates (Shackleton et al., 1993) indicate that an isotope record may be determinable with paleomagnetics and perhaps tephrochonology as additional means of age control.
The Campbell Plateau transect will yield a terrigenous silt record of late Miocene to Holocene fluctuations in current strength near the AAIW/thermocline transition (Site SWPAC-1A), a mainly carbonate record of the Neogene AAIW paleohydrography of the Campbell Plateau (Site SWPAC 6B), and a mixed carbonate abyssal mud record of Neogene paleohydrography of the DWBC (Site SWPAC-7A).
3. Linking Latitudinal Transect (C)
This transect spans a latitudinal range from 51°S to 39°S, commencing on the Campbell Drift (Site SWPAC-7A, Campbell Plateau transect) and passing through the North Chatham Drift (Site SWPAC-2A, Chatham Rise transect). The addition of two other sites completes the transect: Site SWPAC-8A near 47°S on the north levee of the Bounty Fan and Site SWPAC-9A at 39°S on the Rekohu Drift.
The latitudinal transect has been inserted to better track movements of fronts (STC, SAF, and possibly the AAC) during glacial and interglacial cycles. Penetration is therefore limited to the Plio Pleistocene (last 3 m.y.) at Site SWPAC-8A. This site is also deliberately located on a fan levee to retrieve information on sedimentary processes, particularly on the frequency of turbidity currents and their potential as a proxy for paleoseismic events, and the glacial/interglacial switching of sediment supply into the DWBC system.