CONCLUSIONS

ODP Leg 178 had two principal objectives. The first was to test the viability of drilling the glacial prograded wedge on the continental shelf and the derived sediment drifts on the upper continental rise to obtain a record of Antarctic glaciation. The second was to examine Antarctic Peninsula glacial history over the past 8-10 m.y., a period that included both the "warm Pliocene," proposed by some as having seen substantial deglaciation, and the growth and eventual dominance of the Northern Hemisphere ice sheets. These objectives were its particular contributions to an ANTOSTRAT plan to determine a complete Antarctic glacial history by examining similar sediments in several sectors of the Antarctic margin and combining the results by means of numerical models of glaciation.

The drilling succeeded on both counts, demonstrating the validity of the strategy (while determining its limitations) and obtaining a high-resolution record of Antarctic Peninsula glacial history over the past 10 m.y. In more detail, and mainly as a result of drilling and postcruise work on samples and data, we can conclude the following:

1. Recovery in unconsolidated glacial sediments on the continental shelf was very poor, probably because indurated clasts could not be supported by an unconsolidated fine-grained matrix during rotary drilling. Recovery improved at greater depth but was still poor, probably because ocean swell (perhaps incompletely compensated at times) and the necessary use of crude but robust drilling tools prevented fine control of weight-on-bit. Improved recovery might result from drilling tool developments, better heave compensation, and drilling during periods of reduced ocean swell, with compensatory use of logging while drilling. Despite the limitations, shelf sediments can provide a useful record of glaciation; a key lesson of Leg 178 is to ensure that only low-resolution questions are asked of ODP-style drilling on the continental shelf.
2. Leg 178 shelf drilling recovered Pliocene-Pleistocene diamicts, reflecting a subglacial environment (compatible with regular grounding line advance to the continental shelf edge) and corresponding to topset beds of seismic reflection sequence groups S1 and S2.
3. Shelf seismic sequence group S3 is also glacial, but a wider range of environments (subglacial, proglacial, and glaciomarine) is represented in recovered sediments. Recovered sequence group S3 (its lower part was not sampled) is of earliest Pliocene and late Miocene age (back to ~7.5 Ma) and is considered to represent paleoshelf (topset) deposition. A decompaction model, supported by the abundances of benthic diatoms on both shelf and rise, suggests that the sequence group S3-age continental shelf was not overdeepened as it is today.
4. Drilling on the continental rise sediment drifts demonstrated clearly that they hold a high-resolution, recoverable record of continental glaciation. During Leg 178 only their upper part (back to ~10 Ma) was sampled, but with excellent recovery. Sediments are mainly fine-grained alternations of interglacial bioturbated biosiliceous silts and clays and glacial laminated, largely barren terrigenous silts and clays. IRD is present throughout.
5. IRD at the rise sites shows that the ice sheet was present throughout the past 10 m.y., and variations in clay mineralogy indicate that the ice sheet grounding line migrated regularly to the continental shelf edge through that period, including the "warm" early Pliocene and the time of deposition of the sampled part of sequence group S3. Biogenic opal variation suggests a cool late Miocene, a warm early Pliocene, and late Pliocene cooling to a cold (as at the present day) Pleistocene, changes that coincide with global climate changes. Thus, the Antarctic Peninsula ice sheet persisted through the "warm" Pliocene, its volume decoupled from longer-term global climate change and sufficiently great for regular grounding line migration to the shelf edge. The possibility of decoupling from global climate of volume changes of the entire Antarctic ice sheet through this period must also be considered.
6. The opal evidence of correspondence between global climate and southwest Pacific oceanic (Antarctic Peninsula offshore) climate casts doubt on the shipboard interpretation of the upper Miocene shelf sequence group S3 as reflecting "less glacial" conditions than the overlying Pliocene sequence group S2. The change in range of recovered facies might have resulted from improved recovery of more consolidated sediments or from changes in conditions of deposition and preservation less simply related to climate.
7. Regular pre-late Pliocene grounding line advance to the Antarctic Peninsula continental shelf edge shows a sensitivity to sea level change that most probably was caused by volume changes of the main Antarctic ice sheet, since there were then no large Northern Hemisphere ice sheets. Fundamentally, the Antarctic Peninsula ice sheet appears to have acted as a recorder rather than a driver of change, both after and before the development of Northern Hemisphere ice sheets.
8. Spectral analysis of several properties of rise drift sediments (at Sites 1095 and 1096) failed to show the dominance of periodicities corresponding to those of orbital insolation change. The cause of this lack of correspondence is unknown; it could be nonlinear deposition, the inadequacy of the available stratigraphic control, truly decoupled (autocyclic) ice sheet behavior, or a combination of these. The only evidence of sensitivity to orbital insolation change came from IRD, carbonate, and magnetic susceptibility variation at Site 1101 over the past 1.9 m.y. and from correlation between drift sites over the last 200-300 k.y., periods when Northern Hemisphere ice sheet variation drove sea level change. Further investigation of orbital sensitivity is essential, particularly for older periods, and should attempt to distinguish between "oceanic" and "ice sheet" influences, which may have behaved differently.
9. Sedimentation rates on the rise were greater during the warm early Pliocene than during the subsequent, colder late Pliocene and Pleistocene, supporting views that a warmer glacial regime has a greater ice and sediment throughput and suggesting the possibility of sediment starvation around Antarctica should cooling persist.
10. Glacial onset (assumed to be defined as glacial ice grounded significantly below sea level) was earlier than the drilled section. Based on a combination of onshore and offshore data, we suggest an Antarctic Peninsula climate history that included Oligocene (and brief earliest Miocene) glaciation, then was warmer (perhaps nonglacial) until the middle Miocene, then involved renewed glaciation, which has persisted and strengthened to the present day. Such a history would be broadly compatible with global oxygen isotopic ratios measured on benthic foraminifers. An earlier onset than predicted by the glaciological model may have resulted from higher-than-average regional precipitation and from ocean circulation within the Weddell Sea.

In addition, it seems clear that the potential of the rise drift drilling is not yet exhausted. In particular, a wealth of sediment properties measurements is now available to assist further analysis. Some are likely to reflect oceanic influence, others those of the ice sheet. Magnetobiostratigraphic constraints on drift sediment ages have improved and, with possible use of techniques of spectral analysis independent of sedimentation rate, may allow an understanding of the detailed behavior of the Antarctic Peninsula and Antarctic ice sheets over the past 10 m.y. or so.