DISCUSSION

Linear interpolation between inclination reversals recorded in Core 202-1237B-31H (Fig. F4) and the base of Chron 9n in Core 32H (Shipboard Scientific Party, 2003) is used to develop the age model. Calibration to the new GTS 2004 (Gradstein et al., 2004) indicates an age span from ~27.2 to 25.3 Ma and sedimentation rates ranging from 4.05 to 6.79 m/m.y.

Magnetic polarity stratigraphy and oxygen isotope data from Core 202-1237B-31H confirm some previously suggested correlations. In particular, high C. mundulus ¹⁸O values in upper Chron 9n are considered to define Oi2b, the last in a series of distinct ¹⁸O maxima in the Oligocene (Miller et al., 1991). This association has been found at Deep Sea Drilling Project (DSDP) Site 529 (Miller et al., 1991), whereas Oi2b is recorded in Chron 8r at Site 522 (Miller et al., 1988). This ¹⁸O maximum may also correlate with glacial marine sequences in the Ross Sea (Leckie and Webb, 1986). Here we confirm that the highest values interpreted to represent Oi2b are recorded in upper Chron 9n (0.30–0.60) after the late Oligocene ¹⁸O decrease. The initial C. mundulus ¹⁸O decrease at ~26.7 Ma is associated with the last occurrence of P. opima (P21/P22 planktonic foraminifer zonal boundary) between 322.51 and 322.76 mcd in lowermost Chron 8r. This zonal boundary may represent a useful marker for the initiation of the Chron 8r ¹⁸O shift.

The transition from Oi2b to the late Oligocene climate optimum appears to span nearly all of Chron 8r (27.027–26.554 Ma in GTS 2004) (Fig. F5). This transition is composed of three temporary decreases of 0.3–0.5 at ~26.7, 26.5, and 26.35 Ma based on an assumed constant sedimentation rate of 5.52 m/m.y. in Chron 8r. The final decrease to low C. mundulus ¹⁸O values occurs in upper Chron 8r and marks the initiation of the late Oligocene climatic optimum and is here termed the "Chron 8r ¹⁸O shift." The amplitude of our ¹⁸O variations is somewhat less than predicted by recent reconstructions of sea level variability in the Oligocene (Pekar and Miller, 1996; Kominz and Pekar, 2001; Pekar et al., 2002), but we have not estimated the influence of deepwater temperature variability on our ¹⁸O record.

Despite the lack of independent deepwater temperature proxy data, our ¹⁸O record can be used to test the idea that a significant reversal of the long-term "greenhouse to icehouse" trend was initiated by rapid deep-sea warming during the late Oligocene (Zachos et al., 2001). Understanding this reversal in the Cenozoic greenhouse to icehouse trend has important implications for the stability of polar ice sheets. In particular, if ice mass equivalent to the modern-day Antarctic ice disappeared at this time, it might strengthen the case for an unstable East Antarctic Ice Sheet; however, our data and those of Lear et al. (2004) indicate that the late Oligocene ¹⁸O decrease was much less than that reconstructed based on spliced records from different ocean basins (Zachos et al., 2001). Modest changes in Antarctic ice volume are more in line with evidence for persistent grounded ice in the Ross Sea during the late Oligocene (Hayes and Frakes, 1975; Leckie and Webb, 1986; DeSantis et al., 1995) and terrestrial evidence that indicates temperate climates throughout the late Oligocene (Askin, 1992).

Our stable isotope data resolution (~10–20 cm sampling) is not sufficient to define the full spectrum of orbital-scale variability. The ¹⁸O values hint at substantial glacial–interglacial cycles of ~0.6 amplitude, particularly from ~26.7 to 26.3 Ma during the Chron 8r ¹⁸O shift (Fig. F5). Visual inspection shows three distinct episodes in <400 k.y. that may reflect climate cycles in the eccentricity band. In addition, percent coarse fraction data (5 cm sampling) exhibit considerable variability in the 100- to 400-k.y. band during the ~27.2 to 26.3 Ma interval, indicating a promising interval for further study. Percent coarse fraction is controlled by a combination of carbonate production and preservation, including corrosivity to carbonate due to deep ocean ventilation and circulation changes. Our ¹³C data, another index of deep ocean ventilation and circulation changes, do not exhibit a clear relation to ¹⁸O at the current resolution.

Our late Oligocene C. mundulus ¹³C data can be compared with similar existing records from the North Atlantic, Southern, and Indian Oceans to investigate large-scale circulation patterns. The long-term variations in deepwater dissolved inorganic carbon ¹³C recorded by benthic foraminifers are controlled by reservoir changes in carbon cycling, mainly the relative proportion of carbon deposited as organic and inorganic carbon; however, interocean ¹³C gradients at a given time reflect aging of deep waters as they accumulate remineralization products along the flow path (Broecker and Peng, 1982). This approach has been used by many other workers investigating Paleogene deep circulation (e.g., Miller and Fairbanks, 1985; Wright and Miller, 1993). Accordingly, Cibicidoides ¹³C records from these regions were placed on a common timescale by an integrated paleomagnetic and oxygen isotope stratigraphy (Zachos et al., 1996, 2001) and updated to the GTS 2004 (Gradstein et al., 2004).

Examination of Cibicidoides ¹³C records from North Atlantic DSDP Site 563 (33°38.53´N, 43°46.04´W; water depth = 3786 m), Southern Ocean ODP Site 689 (64°31.01´S, 03°05.99´E; water depth = 2200 m), Indian Ocean ODP Site 744 (61°34.66´S, 80°35.46´E; water depth = 2307 m), and Southeast Pacific Ocean Site 1237 reveals interocean gradients very different from the modern ocean (Fig. F6). These records have been correlated to Site 1237 based on magnetic polarity stratigraphy updated to GTS 2004. Interocean ¹³C gradients are small during the late Oligocene (<0.5) Throughout most of the interval studied, however, the highest ¹³C values (youngest deep waters) are recorded at Indian Ocean Site 744 and Pacific Ocean Site 1237, while the lowest values (oldest deep waters) are recorded at North Atlantic Site 563 with intermediate values at Southern Ocean Site 689. Such a pattern is consistent with an Indo-Pacific to Atlantic aging of deep water during the late Oligocene, as suggested by earlier work (Kennett and Stott, 1990). No major change in this pattern is observed in association with the late Oligocene climate transition, but the uncertainties of interocean correlation based on paleomagnetic data with different stratigraphic resolution require caution in interpreting this finding.