The reasons why the Pacific margin of the Antarctic Peninsula was drilled (during Ocean Drilling Program [ODP] Leg 178) have been described in detail by Barker and Camerlenghi (1999). Briefly, the drilling proposal on which the leg was based was one of a linked series (constructed by ANTOSTRAT Regional Working Groups and refined by an ODP Detailed Planning Group; see also Barker et al., 1998, 1999) intended to elucidate Antarctic glacial history, which otherwise is known only from sparse onshore outcrop and existing offshore drilling (for a review, see Barrett, 1996) and from ambiguous and conflicting low-latitude proxy measurements. ANTOSTRAT proposed drilling at several locations around the Antarctic margin, combining the results by means of a numerical model of glaciation (Huybrechts, 1992, 1993). The Antarctic Peninsula's glacial history was thought to have been shorter than that of East Antarctica, and its ice volume would never have been large. Nevertheless, it was the first region of the Antarctic margin to be drilled of those proposed by ANTOSTRAT. It was selected because of the simplicity of its geology, because its narrow ice catchment and anomalously high precipitation would have given rise to a high-resolution glacial record (covering perhaps the Pliocene-Pleistocene and late Miocene), and because (partly on account of the wealth of available marine seismic reflection data) the main features of the region were known and understood. This last was important because of the need to test the strategy of extracting Antarctic glacial history (and therefore the histories of glacioeustatic sea level change and of ice-volume change of oceanic isotopic composition) by direct drilling of glacially transported sediments before drilling in regions with a longer or more complicated glacial history.
The focus of this paper is on developments since publication of the Leg 178 Initial Reports volume (Barker, Camerlenghi, Acton, et al., 1999). A second leg of those proposed by ANTOSTRAT to address the question of Antarctic glacial history has been drilled meanwhile (O'Brien, Cooper, Richter, et al., 2001), in Prydz Bay, East Antarctica (Fig. F1), and related proposals for drilling off Wilkes Land and in the eastern Ross Sea remain under consideration by the advisory structure for ocean drilling. Glacially transported marine sediments have been drilled inshore in the western Ross Sea by the Cape Roberts Project (Cape Roberts Science Team, 1998, 1999, 2000).
Publication in the ODP Scientific Results volume is made available to the shipboard scientific party and others working on samples and data from the drilling leg, but the volume is intended to be flexible so as to respond to whatever role it is given. Many Leg 178 workers have used it to publish interim results, intending to undertake further work or to combine data sets for subsequent external publication. Indeed, many have submitted data reports, within which interpretation and speculation are considered inappropriate. Also, of course, workers have not had access to essential data at a sufficiently early stage (for example, the stratigraphic synthesis and the revised calculation of mean composite depths for two of the continental rise sites could not be made available to those working on rise samples). This paper attempts to synthesize our understanding of Antarctic Peninsula glacial history as it now stands as a result of the papers and data reports published here and taking into account the external literature. Also, however, recognizing the interim nature of some of the work, we include comments on the potential value of studies on Leg 178 samples and data not yet accomplished or completed, which might be out of place in a final synthesis but are justified here.
The Antarctic Peninsula forms a long, narrow dissected plateau extending southwest from ~63°S and merging into West Antarctica at ~74°S. It gives the appearance of an elevated peneplain at ~900 m at the northern end, rising to 1750 m at ~65°S and remaining at or above that level farther south (Elliot, 1997). Its crest is wider in the south but is dissected everywhere by steep-sided fjords that reach well below sea level, and many offshore islands are found on its western (Pacific) continental shelf. That shelf is typical of most glaciated shelves, in being deeper than low-latitude continental shelves and having a reverse (inward) slope. Within the inshore fjords, water depths commonly reach 1000 m and can exceed 1400 m, whereas depths are typically 300-500 m at the continental shelf edge. The continental slope on the western side is abnormally steep (up to 17°).
The elevated Antarctic Peninsula acts as a major barrier to tropospheric circulation and currently receives a relatively high snowfall, almost four times the continental average (Reynolds, 1981; Drewry and Morris, 1992). Over most of the Antarctic Peninsula the climate is fully polar; modern meltwater-influenced sedimentation is known only from the offshore South Shetland Islands in the far north (e.g., Yang and Harwood, 1997). Yet the permanent ice cover on the peninsula spine is thin, and ice sheet drainage occurs mainly through steep ice falls at the heads of fjords, with grounding lines very far inshore and only rare fringing ice shelves. There is considerable evidence, both direct and indirect, that a grounded ice sheet extended to the continental shelf edge at glacial maxima (e.g., Larter and Barker, 1989; Pudsey et al., 1994; Bart and Anderson, 1995; Camerlenghi et al., 1997b; Barker, Camerlenghi, Acton, et al., 1999). In short, the Antarctic Peninsula seems to have had a small-catchment, small-reservoir, high-throughput glacial regime that should provide a high-resolution climate record.
The Antarctic Peninsula has been the site of subduction of Pacific oceanic lithosphere for at least 150 m.y. and probably since long before the breakup of Gondwana. Its onshore geology, comprising magmatic arc and related volcanic products and deformed and metamorphosed sediments of a series of accretionary prisms (e.g., Dalziel, 1984; Barker et al., 1991; Leat et al., 1995) attests to this history. Outcrop along the spine and on islands along the dissected and overdeepened inner shelf is mostly of plutonic rocks. Part of the middle shelf of the Pacific margin is occupied by a sedimentary basin, thought to have been the upper-slope basin of a subducting margin. A mid-shelf high (MSH) at the outer edge of this basin (Fig. F2) is considered to have been a mid-slope high, as seen on many subducting margins, and possibly to represent the edge of the rigid overriding plate. The present outer shelf is most probably underlain at depth by accretionary prism material.
Subduction stopped during the Cenozoic with the arrival of a spreading ridge crest at the trench, progressively later northeastward along the margin (Herron and Tucholke, 1976; Barker, 1982; Larter and Barker, 1991a). For example, ridge-crest collision occurred at ~30 Ma south of 67°S, at 3-6 Ma at ~62.5°S, and at a series of intermediate ages in between (see fig. F7 of Barker and Camerlenghi, 1999). North of 62.5°S, where lies the 4700-m-deep South Shetland Trench and an active backarc extensional Bransfield Strait (Br. St. in Fig. F2) (see also Lawver et al., 1995; Barker and Austin, 1998), spreading stopped before the ridge crest reached the margin but subduction probably continues at the trench (Barker and Dalziel, 1983; Maldonado et al., 1994; Kim et al., 1995).
The ridge-crest collision had several significant effects. Subduction-related magmatism ceased some time before collision (Barker, 1982), but a recent phase of alkalic volcanism has been related by some to creation of a "slab window" following collision (e.g., Hole and Larter, 1993), though the age and location of some occurrences do not fit a simple story (Barber et al., 1991). Oceanic magnetic lineations young toward the margin, indicating collision, but a magnetic quiet zone close to the margin probably signifies an abundance of terrigenous sediment prior to collision, sufficient to bury the approaching ridge crest. Deep Sea Drilling Project (DSDP) Site 325 (Hollister, Craddock, et al., 1976), in 3748 m water depth and ~180 km from the continental shelf edge (Fig. F2), shows a loss of terrigenous sediment from 1 to 7 m.y. after collision (Hollister, Craddock, et al., 1976; D. Lazarus, M. Iwai, L. Osterman, and D. Winter, pers. comm., 2000), which has been interpreted as sediment deflection caused by uplift of the MSH, thermal (Larter and Barker, 1991a) and possibly also hydrothermal (Larter et al., 1997) in origin, followed by passive-margin-like subsidence. The occurrence of the same process (collision followed by temporary MSH uplift and consequent deflection of terrigenous sediment) along the margin has been used to constrain ages of shelf and rise sediments.
ODP Leg 178 drill sites on the continental shelf and rise were chosen to avoid interference between tectonics (collision and uplift) and climate change—the strategy was to sample younger units in the northeast and older units in the southwest. Site 1103 was an exception to this, but its occupation was forced upon the shipboard party as a result of poor penetration and recovery at other shelf transect sites. Drilling was planned also to avoid the young extensional zone in Bransfield Strait. It was thought at one time that the South Shetland margin (including Bransfield Strait) provided a useful example of the margin before collision, but such a comparison neglects changes in climate with time and the backarc extension in Bransfield Strait, which is not mirrored farther south.
The other significant feature of the Antarctic Peninsula is the elevation of its spine. If that spine is a peneplain, then uplift from close to sea level is implied. Such uplift has not been dated, and several hypotheses have been erected to explain it. It is not simply associated with ridge-crest subduction. A relation is inferred by some between uplift and climate change via the development of a topographic barrier to atmospheric circulation (see Elliot, 1997), so that the time of uplift is important to climate history.