RESULTS

We have constructed an age model for Hole 1017E based on a core-top age of 0 ka, using eight ¹⁴C datums (Table 1, Table 2) and 13 oxygen isotopic datums (Table 2). Radiocarbon ages together with their converted calendar-year ages are shown in Table 1. Of the 13 samples analyzed, Sample 167-1017E-1H-3, 33-36 cm, gives a much older ¹⁴C age relative to adjacent samples. Because this level corresponds to a thin turbidite layer, this older age is considered to be the result of reworking; thus, this sample was rejected in constructing the age model. Of the remaining samples, eight were selected to construct the age model for the interval shallower than 7.45 m (Fig. 2) and represent the general trend of the age-depth curve (Fig. 2). The four other radiocarbon datums were not used to construct the age model because they deviate slightly from the curve (up to 400 yr for specific depths). Sediment ages between the datums are estimated by linear interpolation. The radiocarbon age data indicate near-linear sedimentation rates during the last 33 k.y. with an average rate of 18 cm/k.y.

High-resolution benthic oxygen isotopic data for Hole 1017E are listed in Table 3 and plotted against depth in Figure 3. The data exhibit the characteristic sawtooth pattern of the latest Quaternary ¹⁸O deep-sea record from MIS 6 (~130 ka) to the present day (Fig. 4). The isotopic curve is similar to those from the deep sea (Mix, 1987; Martinson et al., 1987) and presents an essentially continuous record from the latest part of MIS 6 (130 ka to present day). The sequence includes two glacial terminations (Terminations I and II), glacial Stages 4 and 2, two interglacial episodes (MISs 5 and 1), and interstadial Stage 3. Also clearly included are the five substages of interglacial Stage 5 (5E through 5A; Fig. 3). The bottom of Hole 1017E coincides with Termination II, which clearly is not complete because the ¹⁸O values do not reflect glacial maxima values, and the shift at Termination II is only 1.1. This compares with the shift of 1.7 associated with Termination I (Fig. 3). Correlation of the gamma-ray attenuation porosity evaluator records between the several holes of Site 1017 revealed two short gaps in the record (Fig. 3) that coincide with the core breaks (Lyle, Koizumi, Richter, et al., 1997).

Unambiguous paleoclimatic events recorded in Hole 1017E were correlated with the standard deep-sea oxygen isotope chronology (Martinson et al., 1987). Thirteen isotopic events are recognized and shown in Table 2 with assigned ages. The benthic oxygen isotopic record has been dated using these events (Fig. 4). Sediment ages between the events were linearly interpolated. The near-linear sedimentation rates for the upper part of the core continue downward for the remainder of the sequence (Fig. 2). Average rates of sedimentation in the middle and lower parts of the sequence (130-33 ka) are 17.2 cm/k.y.

The chronological resolution of the benthic oxygen isotopic data in Hole 1017E is every 334 yr. Each sample (3 cm thickness) analyzed represents an interval of 167 yr.

To help develop a stratigraphic framework for Hole 1017E, we established a high-resolution planktonic foraminiferal ¹⁸O record for G. bulloides for the last 60 k.y. (Fig. 5). This has provided an exquisite record of climate change including the Holocene, Younger Dryas, Bølling/Ållerod (Dansgaard/Oeschger [D/O] 1), last glacial maximum, and D/O Cycles 17 through 4 (Fig. 5). The beginning of the Bølling/Ållerod is clearly marked by a rapid ¹⁸O decrease of 2.5. The D/O cycles are represented by distinct ¹⁸O changes ranging from a few tenths per mil to 1.2. As in the Santa Barbara Basin (Hendy and Kennett, 1999), the planktonic foraminiferal ¹⁸O record in Hole 1017E is considered to primarily represent regional changes in sea-surface temperature instead of salinity. The D/O cycles in Hole 1017E have been identified by visual comparison of the ¹⁸O with the ¹⁸O sequence recorded in the Greenland Ice-Core Project (GRIP; Fig. 5). Sufficient differences occur in the character of individual D/O cycles to allow unequivocal identification of many of the cycles (including D/O Events 17, 14, 13, 12, and 8).

We experienced the greatest difficulty identifying D/O Cycles 6 through 4 because these are of shorter duration; thus, their signals are attenuated. D/O Event 4 was identified from its associated radiocarbon date. A relatively cool stadial separates and assists with the identification of Events 5 and 4. The sediment sequence is incomplete in the interval immediately following D/O 4, associated with the core break, and it appears that D/O 3 is missing. A conspicuous offset occurs in ages of the sequence of D/O cycles between the Greenland ice core and Hole 1017E (Fig. 5). Each D/O cycle is shown to be younger in Hole 1017E compared with the Greenland ice core. The reason for this relatively systematic age offset is unclear. We believe that these climatic events are almost certainly isochronous (Hendy and Kennett, 1999), and therefore the age offset is an artifact of the time scale. This is difficult to prove given the strong limitation of the radiocarbon time scale (Voelker et al., 1998) and insufficient resolution in using SPECMAP age assignments for the sequence older than 35 ka.

To better compare the climatic records from Hole 1017E and GRIP, we have tied them together by correlating three intervals representing the initiations of interstadials 14, 12, and 8 (Fig. 6). The effect of this is to bring the sequence of D/O cycles into alignment. Although the correlations are made using only three tie points, strong similarities are revealed in the timing and magnitude of climate patterns between the two records. An effect of this chronological realignment is to reassign the age for the level of the oldest radiocarbon date (32,958 ± 380 ka) in the sequence to ~36,500 ka. Strong deviations in radiocarbon chronology are known for this interval (Voelker et al., 1998).

Termination IA is shown by a rapid ¹⁸O decrease of 2.5 in G. bulloides and Termination IB by a decrease of 1.7. During the deglacial interval, a temporary increase of 1 marks the Younger Dryas cool episode (Fig. 5).

The recognition of an almost complete suite of D/O cycles provides a high-resolution stratigraphic framework that will be of value in correlating paleoenvironmental changes recorded in Hole 1017E with other sequences. Paleoceanographic interpretations based on the stable isotopic record of Hole 1017E are to be discussed elsewhere.