DISCUSSION

Interpretation of the IRD record and lithofacies in the paper suggests that there are eight glacial-interglacial cycles from 95 to 38 mbsf (1.37-0.54 Ma). This cyclicity is best recorded in the MS data because of the reduction in MS resulting from abundant carbonate in the interglacial periods (Fig. F3). From 95 to 67 mbsf, the interglacial and glacial intervals are thin relative to those above. Estimates of their duration using linear sedimentation rates indicate that these shorter-period cycles closely match the 41-k.y. cycles of orbital obliquity reported from marine cores in the Northern Hemisphere (Ruddiman et al., 1986; Ruddiman and Raymo, 1988). At ~1.01 Ma (72.4 mbsf), the glacial and interglacial intervals become thicker and the period lengthens to about 100 k.y. The timing of this transition is slightly earlier than that reported in the Northern Hemisphere (Ruddiman et al., 1986). At Site 1101, there is one long cycle prior to the Brunhes/Matuyama boundary; in the Northern Hemisphere, the 41-k.y. cycles continue up to 0.73 Ma.

Ice rafting follows a regular pattern throughout each glacial and interglacial period, with low or no IRD MAR in the first part of the glacial period and an increase after the midpoint with peaks >0.05 g/cm2/ k.y. (Fig. F3). IRD MAR remains high within the interglacials and returns to low values prior to the start of the next glacial period. This pattern of ice rafting is consistent with rapid disintegration of ice streams and release of icebergs from the continental shelf. The IRD MAR peaks within the laminated fine-grained glacial units indicate very high rates of deposition, because the overall sedimentation rates were high during that time. During interglacials, the sedimentation rates were slower and allowed bioturbation, which would have concentrated the coarse-grained IRD within the biogenic-rich sediment. Significantly more debris was deposited in the peaks that occur around 0.88 Ma (between 62 and 60 mbsf) (Fig. F3) than at other times. This could result from extreme deglaciation of the Antarctic Peninsula that exposed debris-rich basal ice from glaciers draining inland fjords (cf. Anderson and Andrews, 1999). This rafting event occurs within the first 100-k.y. glacial interval following the transition from 41-k.y. to 100-k.y. cyclicity. Following this large peak, IRD MAR is lower, with only one peak >0.05 g/cm2/k.y. near the start of the next glacial. The previous pattern of ice rafting does not appear to be reestablished immediately after this large peak (Fig. F2B, interval c).

Lithofacies differ between glacial periods, but they are dominated by thickly laminated mud corresponding to deposition of either distal turbidites or plumites. Thinly laminated mud also occurs regularly within 41-k.y. glacial cycles below 69 mbsf and suggests that periods occurred when turbidity current activity was less and contour currents caused reworking. Within the 100-k.y. glacial intervals, thickly laminated mud alternates with homogeneous mud (Fig. F3). For example, above 53 mbsf there are six fining-upward packages of laminated mud grading into homogeneous mud that are easily observed on X-radiographs because of the density contrast between these two lithofacies. These may represent shorter-period cycles of glacier advance and retreat between proximal and distal locations on the shelf that did not occur previously.

IRD has been transported to the continental rise of the Antarctic Peninsula cyclically over the past 3 m.y. The first prominent peak at Site 1101 occurs at 2.8 Ma (interval a, Fig. F2B) prior to the onset of widespread Northern Hemisphere glaciation at ~2.4 Ma. Generally warm conditions are proposed from 2.9 to 2.48 Ma in the Southern Ocean (Ciesielski et al., 1982). Warmer glacial conditions are suggested at Site 1101 by the presence of diatom ooze within the laminated sediments and the highest sedimentation rates over the last 3.0 m.y. (Barker, Camerlenghi, Acton, et al., 1999). At Site 1101, intense ice rafting at ~2.8 Ma deposited diamicton beds associated with the large IRD MAR peak. However, IRD MAR is low at sites further away from the Antarctic continent, such as in the Argentine Basin (Bornhold, 1983), which may indicate fewer far-traveled icebergs because of more rapid iceberg melt closer to the continent. Warmer conditions on the Antarctic Peninsula may have resulted in warm-based glaciers that calved debris-laden icebergs such as those from Greenland, which would not travel as far as large tabular Antarctic icebergs.

Well-defined 41-k.y. cyclicity occurs at Site 1101 from 1.9 to 1.01 Ma. Prior to this time, IRD MAR peaks are of variable frequency and there are periods without peaks (i.e., 2.12-1.98 Ma). The onset of 41-k.y. cyclicity coincides with prominent peak b (Fig. F2B). The period between 2.6 and 1.67 Ma has been described as an intermediate stage in Antarctic cryosphere development (Warnke et al., 1992). The largest peak at 0.88 Ma (Fig. F2B, interval a) is probably related to the mid-Pleistocene climate transition dated from 0.92 to 0.90 Ma in the Northern Hemisphere (Berger and Jansen, 1994). IRD is at a maximum during this time in the Southern Ocean (Ledbetter and Watkins, 1978) and the Argentine Basin (Bornhold, 1983). Bornhold (1983) attributed high IRD MAR from 0.90 to 0.65 Ma to changes in ocean circulation and more melting rather than increased Antarctic glaciation. Ice rafting from the Scandinavian ice sheet was also high in the Norwegian Sea during the mid-Pleistocene climate transition (Berger and Jansen, 1994). This bipolar connection between ice-rafting events may be related to modification of deep-ocean circulation. The North Atlantic Deep Water feeds the Antarctic Circumpolar Current, thus linking the North Atlantic with the Antarctic continent. Site 1101 is much closer to the continent than these other sites and therefore may be more influenced by the extent of deglaciation of the continent rather than ocean circulation. However, the widespread ice-rafting event at ~0.90 Ma may record changes on the continent and in ocean circulation so that icebergs could deposit debris both at proximal sites and in distant oceans. At Site 1101, prominent peak a (Fig. F2B) also coincides with the transition from 41-k.y. to 100-k.y cyclicity.

NEXT