Early Palaeogene planktic foraminiferal and carbon isotope stratigraphy, Hole 762C, Exmouth Plateau, northwest Australian margin

Although the northwest margin of Australia is an important region for petroleum exploration and palaeoceanographic investigations, its Palaeogene stratigraphy is poorly documented, especially in terms of a foraminiferal biozonation. Early Palaeogene cores from 502.96 to 307.80 m below sea floor at Ocean Drilling Program Site 762 on the Exmouth Plateau were examined in this study for their planktic foraminiferal assemblages and the carbon isotopic compositions of Subbotina spp. Planktic foraminifera are generally well preserved and belong to 74 species and 17 genera. In spite of a mid-latitudinal palaeolatitude (c. 40°S) the sequence, deposited between the early Paleocene and Middle Eocene, contains all planktic foraminiferal Zones P1c through P10 of the current global scheme for tropical locations, except for Subzone P4b. Most zones are well defined by the datums of primary marker species except P3a and P9, which have boundaries that probably occur in core gaps, and the P9 zonal boundaries are defined by secondary marker species. Overall, variations in δ13C based on sequential samples of Subbotina are similar in pattern and magnitude to global summary isotope curves spanning the early Palaeogene. However, the prominent δ13C excursion that characterizes the Palaeocene/Eocene transition is mostly missing and appears to lie in a core gap. The planktic foraminiferal zonation, linked with that based on nannofossils, a recalibrated magnetostratigraphy and carbon isotope records, provides a robust temporal framework for the Early Palaeogene of the northwest margin of Australia.


INTRODUCTION
The Early Palaeogene c. 65 to 49 Ma ( Fig. 1) is an especially significant interval of time because of profound changes in climate, ecosystems and the global carbon cycle (e.g. Zachos et al., 1993Zachos et al., , 2001Corfield, 1994;Thomas & Shackleton, 1996;Dickens et al., 1997;Norris & Röhl, 1999). Indeed, documentation of Palaeogene paleoceanography is now a high priority for stratigraphic research. However, to resolve outstanding questions of global change, early Palaeogene marine successions must be located that meet certain criteria. These targets must be: (1) thick so as to maximize the temporal quality of the record and time resolution; (2) buried with little overburden to reduce the logistical demands of drilling and diagenetic effects; (3) rich in well-preserved biogenic carbonate for geochemical evaluation; and (4) hosted in stratal geometries which permit construction of detailed depth and latitudinal transects. To date, one of the few investigated locations which meets these criteria is the recently drilled Blake Nose in the western North Atlantic. Studies of this region have already advanced our knowledge of Palaeogene environmental change significantly (e.g. Bains et al., 1999;Katz et al., 1999;Norris et al., 1999;Norris & Röhl, 1999).
The Exmouth Plateau (Fig. 2) is a submerged platform extending 150-500 km off northwest Australia that consists of thinned and tilted continental crust with a Phanerozoic sedimentary sequence exceeding 10 km (Barber, 1988;Stagg & Colwell, 1994). The stratigraphic succession on the plateau is particularly significant because it contains a somewhat expanded Palaeogene sequence deposited in moderate water depths and within the southern subtropical zone (Exon et al., 1992). Moreover, unlike many other regions in the world, the Exmouth Plateau has extensive seismic coverage because of petroleum exploration. Seismic records indicate a substantial Palaeogene record which includes thick successions in both offshore and inshore locations making the Exmouth Plateau a potential location for detailed Palaeogene oceanographic reconstructions.
Current understanding of Palaeogene stratigraphy for the northwest margin of Australia is mostly a conglomeration of data from sidewall cores and ditch-cuttings largely held outside of the public domain. Ocean Drilling Program (ODP) Sites 762 and 763 were drilled by Leg 122 in 1988 on the Exmouth Plateau in part to provide an accessible stratigraphic framework for the Tertiary and Cretaceous. Although Neogene and Cretaceous intervals are now well documented for these sites (Wonders, 1992;Zachiariasse, 1992), other than the ranges of calcareous nannofossils (Siesser & Bralower, 1992), and analyses of bulk carbonate carbon isotopes (Thomas et al., 1992), the Palaeogene stratigraphic record has been given little attention. Planktic foraminifera have not been rigorously examined, carbon isotopes have not been measured on foraminiferal species or calibrated against other time markers, and the interpretation of polarity chrons (Galbrun, 1992) is inconsistent with the nannofossil stratigraphy (Siesser & Bralower, 1992). Moreover, the timescale used in previous work (Shipboard Scientific Party, 1990) is now outdated, precluding direct age comparisons with other regions. This study addresses these deficiencies by examining planktic foraminifera and their isotopic composition in sediment from ODP Site 762. The new dataset is combined with

Samples
Ocean Drilling Program Hole 762C (Fig. 2) was drilled on the central Exmouth Plateau (19(53.24#S,112(15.24#E) in 1360 m of water depth. Preliminary shipboard work (Shipboard Scientific Party, 1990) and shore-based nannofossil investigations (Siesser & Bralower, 1992) documented 550 m of Middle Eocene to Lower Campanian nannofossil chalks and oozes from the recovered core. This study investigates 83 samples taken from some 200 m of sediment deposited between the late Early Palaeocene and the early to Middle Eocene. Bulk sediment in this interval consists predominantly of white to pale yellow (5Y 8/2) or grey green (10YR 8/1) carbonate ooze and chalk.
Individual samples of c. 20 g were collected at c. 1.5 m intervals from cores 762C-16X-4 to -37X-4 (307.5 to 503.5 m below sea floor (mbsf)). Portions of each sample were processed for microfossils at the Australian Geological Survey Organization (AGSO) palaeontological laboratory. These samples were treated with hydrogen peroxide to remove organic matter, sieved to remove the <63 µm fraction, and stored in glass vials. The residues were picked for planktic foraminifera, either to completion or to obtain a complete record of taxonomic content. At least 200 and upwards of 500 specimens from each sample were separated and examined as representative of the faunal record.

Biostratigraphy
The identity, relative abundance of individual foraminiferal species, and distribution across the sample set from Site 762 were determined by optical microscopy and scanning electron microscope. Relative abundances of species were classified as dominant (D, >30%), common (C, 10 to 30%), rare (R, 2 to 10%) and extremely rare (X, <2%). Most specimens belong to the genus Acarinina (A), Subbotina (S), Morozovella (M), Globanomalina (G), Chiloguembelina (C) or Parasubbotina (P). The preservation state of the tests also was noted as 'poor', 'moderate' or 'good'. Tests with good preservation show minor secondary calcite growth or dissolution. Tests with moderate preservation are often infilled with calcite or fragmented, but can be classified to species. Tests with poor preservation are moderately to heavily encrusted with secondary calcite, and species identification is less certain.
The current planktic foraminiferal zonal scheme by Berggren et al. (1995) is used in this study. Although Site 762 was at southern mid-latitudes (c. 40() during the Palaeogene (Veevers et al., 1991;Clarke & Jenkyns, 1999), the low latitude biozonation scheme by Berggren et al. (1995) best describes the observed faunal assemblages. Palaeocene foraminiferal classification, including ranges and zonal boundaries, has been summarized recently by Olsson et al. (1999). Boundaries between biozones were placed halfway between the sample containing a defining marker species and the sample above or below it that lacks it.

Stable isotopes
Specimens of Subbotina were separated and analysed for their carbon isotope composition. Although Subbotina probably inhabited deep surface waters in the thermocline (Berggren & Norris, 1997) this genus is considered especially appropriate for  Berggren et al., 1995) and the global benthic foraminifera carbon isotope curve (adapted from Hall, 1990 andZachos et al., 2001).  Exon et al., 1992). Bathymetry in metres. such work because carbon isotope composition shows little inter-species variation (Berggren & Norris, 1997) and sizerelated variation (D'Hondt et al., 1994;Norris, 1996). However, as an added precaution to minimize the effect of ontogenetic changes in depth ecology, we restricted the size range of tests (210 to 260 µm diameter) and collected 20-30 specimens for each analysis. Although no single Subbotina species spanned the entire early Palaeogene, species with suspected phylogenetic affinities (Olsson et al., 1999) and similar morphologies within the genus were collected. The species collected for isotope analyses were S. triloculinoides (Zone P1c to the middle of Subzone P4a-b), S. triangularis (Subzone P4a/b to Subzone P6a) and S. patagonica (Subzone P6a to Zone P10).
Stable isotopes were analysed at the isotope laboratory in the Earth Science Board at the University of California, Santa Cruz (see Billups et al., 1998). Samples were subjected to ultrasound whilst bathed in methanol for 3-5 s to remove adhering particles. Samples then were reacted at 90(C in H 3 PO 4 on an autocarb common acid bath with the CO 2 generated then analysed in a Prism gas source mass spectrometer. Carbon isotope values were calibrated to the Peedee belemnite (PDB) standard, and converted to conventional delta notation ( 13 C).

Foraminiferal biostratigraphy
Planktic foraminifera contribute between 10% and 20% of bulk sample volume. The planktic to benthic foraminiferal ratio for the early Palaeogene is approximately 15:1 and the benthic foraminiferal component rarely exceeds 2%. Microfossils are generally complete and unabraded. Secondary recrystallization or dissolution hampers the identification of species in only a few intervals (Fig. 3), notably within the foraminiferal biozones P4 and P7 (discussed below).
Some 74 species belonging to 17 genera were identified at Site 762, and established biozones can be recognized from the pattern of faunal succession (Figs. 3,4). All zones from P1c to P10, with the exception of Subzone P4b, are found at Site 762, spanning the interval from the early Palaeocene (Danian) to the early Middle Eocene (early Lutetian). Absolute age assignments are from Berggren et al. (1995). Boundaries between different zones are well defined unless noted otherwise. The zonal scheme adopted here generally follows that of Berggren et al. (1995) but some modification is required.
P1c. FAD of Globanomalina compressa and/or Praemurica inconstans to FAD of Praemurica uncinata (500.72 mbsf). Age: 63.0-61.2 Ma, early Palaeocene (Danian). Subzone P1c extends below our sampled interval as only two samples are representative of it. The upper boundary of Subzone P1c is difficult to locate in some sequences because of the short stratigraphic range of its marker species, Pr. uncinata (Pl. 1, figs 1-3). However, this species has been found at Site 762, and is a useful correlation tool. Common species found in uppermost Subzone P1c include Pr. pseudoinconstans, Parasubbotina varianta, Subbotina trivialis, S. triloculinoides, Globanomalina compressa, Chiloguembelina midwayensis and C. subtriangularis. Faunal assemblages within this subzone are consistent with the global scheme (Olsson et al., 1999).

P4.
Total range of Globanomalina pseudomenardii (470.75 to 423.03 mbsf). Age: 59.2-55.9 Ma, middle Late Palaeocene (late Selandian-Thanetian). The lower boundary of Zone P4 is placed with moderate conference because Globanomalina pseudomenardii has a patchy occurrence at Site 762 and has a similar morphology to both G. ehrenbergi and G. chapmani (Pl. 1, figs 10-12). Indeed, preliminary work on the foraminifera at Site 762 (Shipboard Scientific Party, 1990) described Globanomalina pseudomenardii as the inflated 'chapmani type'. However, there is morphological variability within this species with some of the smaller specimens displaying a distinct keel and a Early Palaeogene planktic foraminifera sharpened angular periphery. It is one of these more typical Globanomalina pseudomenardii morphotypes that marks the P4/P5 zonal boundary.  (Olsson et al., 1999). However, at Site 762, this datum occurs in the upper part of Subzone P4c. Consequently, Subzones P4a and P4b cannot be distinguished at Site 762, and they have been combined. Subbotina triloculinoides, S. triangularis, S. velascoensis and Chiloguembelina midwayensis. Such an assemblage is characteristic of Subzones P4a and P4b in the global scheme (Olsson et al., 1999).  (Olsson et al., 1999). The interval assigned to this zone at Site 762 is relatively large compared to that of the global scheme (Berggren et al., 1995) and its base in particular does not correlate well with the nannofossil zones for the site (Siesser & Bralower, 1992). However, we regard the lower boundary of Subzone P4c as being well defined at Site 762 because so many species make their first appearances at its base, representing a pattern consistent with that of the global scheme.  3) and may not represent its true range. Such a pattern may relate to the higher latitudes and cooler temperatures for the Exmouth Plateau during the time of deposition (due to its location at that time, well to the south), as the species is typical of more tropical areas (Olsson et al., 1999) (Berggren et al., 1995). The FADs of Morozovella dolobrata and Acarinina wilcoxensis occur simultaneously in upper Zone P5, which correlates well with their FADs in New Zealand near the base of the Waipawan (Hornibrook et al., 1989). The LADs of Morozovella acuta and M. occlusa occur in upper Zone P5, consistent with the global scheme (Olsson et al., 1999). However, the distinct morphotypes Morozovella allisonensis, M. africana and Acarinina sibaiyaensis that developed and diversified during a brief interval of P5 (e.g. Kelly et al., 1998;Pardo et al., 1999) were not found at Site 762. As discussed later, these species may not have been recovered because of a 5 m core gap in uppermost P5 between 762C-28X-1 and -27X-4.  (Berggren et al., 1995) (Blow, 1979, as Subzone Zone P7;Berggren et al., 1995).

P6. LAD of
There is a problem within the interval containing Subzone P6a at Site 762 with the labelling of core -26X. According to the drilling log, the depth of this core is 398.0 to 402.5 mbsf (4.5 m thick). However, there is c. 8 m of core photographed and illustrated in the lithological log for core -26X. As the next coring interval, -27X, starts at 402.5 mbsf, there is c. 3.5 m of core that has not been accounted for.  (Siesser & Bralower, 1992), recalibrated magnetostratigraphy (this study) and carbon isotope data for Subbotina (this study) and the bulk sediment (Thomas et al., 1992) for the early Palaeogene interval at ODP Site 762. Ages follow Berggren et al. (1995). Larger circles with dashed lines are 13 C, Subbotina specimens; small filled circles are 13 C, bulk sediment.

P6b. FADs of
lensiformis and M. formosa coincide, although at Site 762 the latter species is rare. The FAD of Pseudohasterigina wilcoxensis also occurs at the base of Subzone P6b at Site 762. In tropical assemblages this datum occurs at the P5/P6 boundary but has a delayed entry in mid-high latitude regions within the P6b-P7 interval (Blow, 1979, as Subzone Zone P7;Berggren et al., 1995).  (Blow, 1979, as Subzone Subzone P8a; Berggren et al., 1995). However, at other locations, the LAD of M. subbotinae occurs within Subzone P6b (Berggren et al., 1995), but at Site 762, this datum occurs later in Zone P7.  (Blow, 1979, as Subzone 8b;Berggren et al., 1995). The LAD of Morozovella aequa occurs in Zone P7 in the global scheme (Olsson et al., 1999), however, at Site 762, it occurs slightly later in our Zone P8.  (Berggren et al., 1995). This primary marker species is absent at Site 762. Other authors have noted that Planorotalites palmerae has a very patchy geographical distribution (Toumarkine & Luterbacher, 1985). In the absence of Planorotalites palmerae, the top of this zone is defined by the FAD of Acarinina cuneicamerata (Pl. 2, figs 8-10) and the common appearance of A. bullbrooki (W. A. Berggren, pers. comm., 1999). Common species found in this zone at Site 762 and elsewhere (Toumarkine & Luterbacher, 1985) (Toumarkine, 1981;Berggren et al., 1995). However, this datum occurs higher in the sedimentary column at Site 762 (section 762C-15X-2) co-existing with taxa characteristic of Zone P11. The placement of this boundary has been problematic elsewhere because of the late arrival and/or rare early appearance of Hantkenina nuttalli (e.g. McGowran, 1974). The LAD of Planorotalites palmerae is another criterion for the P9/P10 boundary (W. A. Berggren, pers. comm., 1999). Unfortunately, this species was not found at Site 762. The FAD of Guembelitrioides higginsi (Pl. 2, figs 11-13) is close to the P9/P10 boundary at other locations (W. A. Berggren, pers. comm., 1999). We therefore use this datum to mark the top of Zone P9. Common species in Zone P9 at Site 762 and elsewhere (Toumarkine & Luterbacher, 1985;Berggren et al., 1995) (Berggren et al., 1995).  5. Sedimentation rates calculated from planktic foraminifera (this study) and nannofossils (Siesser & Bralower, 1992) compared with published magnetostratigraphy (Galbrun, 1992) and recalibrated magnetostratigraphy (this study). The original magnetostratigraphy deviates from the other curves, showing where the major adjustment was made for Chron 22n at 49 Ma. Sedimentation rates calculated from the planktic foraminiferal and nannofossil biozonation using the global scheme are 1.5 cm ka K1 and 1.4 cm ka K1 respectively. P, planktic foraminiferal biozones (this study).

Carbon isotopes
The 13 C of Subbotina tests averages 1.84‰ across our sampled interval but varies significantly between 0.36 and 2.92‰ (Table 1; Fig. 4). From the lowest sampled interval in P1 (502.96 mbsf) to the base of Zone P5 (423.49 mbsf) 13 C gradually increases by about 2.0‰. There is a marked decrease in 13 C through Zone P5 (423.49 to 407.95 mbsf) by about 2.5‰. From the top of Zone P5 (407.95 mbsf) to the top of the sampled interval in Zone P10 (307.80 mbsf) 13 C is relatively constant, except for a slight drop in upper Zone P6. The Subbotina 13 C curve mirrors the bulk carbonate 13 C curve at Site 762 (Thomas et al., 1992), although Subbotina 13 C values are typically 0.5‰ lower at similar depth. This offset in 13 C is expected because surface waters are generally enriched in 13 C relative to deeper waters of the thermocline.
The exogenic carbon cycle includes all carbon stored in the ocean, atmosphere and biomass reservoirs. Secular changes can occur in this exogenic carbon cycle with variations in carbon inputs or outputs to the ocean or atmosphere (e.g. Kump & Arthur, 1999;Dickens, 2001). Because carbon cycles through all reservoirs of the exogenic carbon cycle over relatively short time intervals, about 2000 years at present day, major perturbations in the isotopic composition of the exogenic carbon cycle will be observed in all carbon reservoirs at nearly the same time (e.g. Dickens et al., 1997). The early Palaeogene is especially amenable for correlating widespread locations by this "carbon isotope stratigraphy" because there are a series of prominent 13 C excursions in the exogenic carbon cycle of long and short duration (e.g. Zachos et al., 1993;Corfield, 1994;Dickens et al., 1997, Zachos et al., 2001. Most key features of the global carbon isotope record for the Palaeogene (Fig. 1) can be recognized at Site 762 (Fig. 4). Importantly, when the 13 C variations at Site 762 are placed within the timescale defined by our foraminiferal biozones, they match within 0.5 Ma the age pattern suggested for the global 13 C record by Berggren et al. (1995). This cross-correlation supports the planktic foraminiferal zonation we have developed for Site 762.

Integrated biostratigraphy
In the previously published nannofossil biostratigraphy for Site 762 (Siesser & Bralower, 1992), nannofossil datums and biozones were calibrated to the timescale of Haq et al. (1987). We have recalibrated the nannofossil biozones with the most current global timescale (Berggren et al., 1995). The revised nannofossil biostratigraphy at Site 762 agrees reasonably well with both the planktic foraminiferal and carbon isotope stratigraphy. However, there are discrepancies, notably the placement of the NP6/NP7, NP8/NP9, boundaries, which appear to be too high in the column (Fig. 4). There are several possible reasons for such anomalies. First, key nannofossil markers are very rare at Site 762 (Shipboard Scientific Party, 1990) and may have been missed. Second, the high latitude of the Exmouth Plateau in the early Palaeogene may have put some species at the limits of their biogeographic range (Wonders, 1992) resulting in certain nannofossil and foraminiferal biomarkers at Site 762 being responsive to distribution patterns which applied to the north Australian margin rather than the temporal controls of evolution and extinction. Third, the mismatches could, at least in part, be artefacts of sampling resolution. Galbrun (1992) determined the pattern of polarity reversals for Upper Cretaceous and Lower Tertiary core obtained from Site 762. As noted by Berggren et al. (1995, p. 184), the magnetostratigraphy raises a correlation problem. Starting with Chron C22n and continuing back through the Palaeocene, there are significant discrepancies between magnetostratigraphic and nannofossil datums at Site 762 (cf. Siesser & Bralower, 1992;Galbrun, 1992). In particular, Chron C22n has been placed within Zone NP12. However, according to global stratigraphic schemes, Chron C22n occurs within NP14 (Berggren et al., 1995). Three plausible explanations could account for the miscorrelation. First, the nannofossil zones at Site 762 could be diachronous by some 1.5 Ma with the global timescale. Second, recognition of key nannofossil zones could be incorrect. Third, the polarity chrons could have been mislabelled or misinterpreted. It is noteworthy that the P7/P8 zonal boundary roughly coincides with the top of Chron C23n in global schemes (Berggren et al., 1995), but at Site 762 it coincides with the chron labelled C22n (Galbrun, 1992). It is therefore likely that Chron C23n was mistaken for Chron C22n in previous work. Making this correction, and relabelling the chron succession as a consequence, results in a magnetostratigraphy which correlates well with both the planktic foraminiferal zonation and the isotope stratigraphy (Figs 4,5).

Palaeocene/Eocene transition
The transition between the Palaeocene and Eocene epochs is characterized by a brief warming event c. 55.5 Ma. Globally, this event is recognized in marine sediment sequences by a pronounced benthic foraminiferal extinction event, the appearance of distinct planktic foraminiferal morphotypes, and a remarkable K2.5 to K3‰ excursion in 13 C (e.g. Kennett & Stott, 1991;Bains et al., 1999). The distinct morphotypes and abrupt 13 C anomaly were not observed at Site 762. The simplest explanation is that the warming event lies in a 5 m core gap in uppermost Zone P5 between 762C-28X-1 and -27X-4. One check on this interpretation is the occurrence of Gavelinella beccariiformis. This species dominates benthic foraminiferal assemblages during the Palaeocene but disappears at the warming event. At Site 762 it last occurs in section -28X-1, immediately below the core gap. A K1.3‰ 13 C excursion is represented in bulk carbonate from section -28X-2 to -28X-1 (Thomas et al., 1992). The start of the warming event may thus be registered in section -28X-1.

Sedimentation rates
Collectively, all biostratigraphic and revised magnetostratigraphic datums show a near-linear age-depth relationship. Subzone P4c has a greater rate of sedimentation, which may relate to local processes such as uplift. However, the overall coherency permits the accurate calculation of sedimentation rates for the early Palaeogene sequence at Site 762. Sedimentation rates are remarkably constant at about 1.5 cm ka K1 , which is typical for carbonate oozes on continental margins (Kennett, 1982) but relatively high compared to known Palaeogene sequences that have not been deeply buried. Continuous sedimentation rates also indicate that significant hiatuses are unlikely to exist at Site 762.  Berggren et al. (1995). Early Palaeogene planktic foraminifera microfossils, including planktic foraminifera. We have constructed the first detailed early Palaeogene planktic foraminiferal biostratigraphy at this site, and for the region in general (Fig. 3, Table 2). Sediment deposited between 307.8 and 502.96 mbsf spans Zones P1c to P10 (Berggren et al., 1995), although Subzone P4b cannot be recognized. Despite a relatively high latitude Palaeogene location for Site 762, planktic foraminiferal biozones are generally in phase and contain representative assemblages with those of the currently used global scheme for sub-tropical locations. However, rare, patchy or non-occurrences of the zonal marker species such as Globanomalina pseudomenardii, Morozovella velascoensis, Morozovella formosa, Planorotalites palmerae and Hantkenina nuttalli, make some correlations difficult. The 13 C record constructed from tests of Subbotina spp. is similar to that constructed from bulk sediment. Both records show broad-scale excursions which match the global 13 C curve. However, the large and short-lived negative 13 C excursion that marks the Palaeocene/Eocene transition is at best partially represented at Site 762, probably because of a core gap. Nannofossil biozonation and magnetic polarity chron records (after adjustment) are consistent with the planktic foraminifera biozonation and the 13 C curve (Table 2). Despite core gaps, the integrated stratigraphy at Site 762 suggests that a complete and expanded Early Palaeogene sediment record exists on the Exmouth Plateau. The relatively shallow subsurface depth of this Palaeogene sequence makes the Exmouth Plateau an ideal location for future scientific drilling to understand Palaeogene oceanography.