The timing and duration of the coldest period in the last glacial stage, often referred to as the last glacial maximum (LGM), has been observed to vary spatially and temporally. In Australia, this period is characterised by colder, and in some places more arid, climates than today. We applied Monte-Carlo change point analysis to all available continuous proxy records covering this period, primarily pollen records, from across Australia (n = 37) to assess this change. We find a significant change point occurred (within uncertainty) at 28.6 ± 2.8 ka in 25 records. We interpret this change as a shift to cooler climates, associated with a widespread decline in biological productivity. An additional change point occurred at 17.7 ± 2.2 ka in 24 records. We interpret this change as a shift towards warmer climates, associated with increased biological productivity. We broadly characterise the period between 28.6 (± 2.8) – 17.7 (± 2.2) ka as an extended period of maximum cooling, with low productivity vegetation that may have occurred as a combined response to reduced temperatures, lower moisture availability and atmospheric CO2. These results have implications for how the spatial and temporal coherence of climate change, in this case during the LGM, can be best interrogated and interpreted.

Geomorphic, stratigraphic, palynologic and 14C evidence indicates that the West Coast Range, Tasmania, was glaciated at least three times during the late Cenozoic. The last or Margaret Glaciation commenced after 30,000 yr B.P., culminated about 19,000 yr B.P., and ended by 10,000 yr B.P. During this period a small ice cap, ca. 250 m thick, and cirque and valley glaciers covered 108 km2. The glacial deposits show little chemical weathering or erosional dissection. The snow line ranged from 690 to 1000 m with an average of 830 m for the ice cap. Mean temperature was 6.5°C below the present temperature. During the preceding Henty Glaciation a 300- to 400-m-thick ice cap and outlet glaciers exceeded 1000 km2. The glacial deposits are beyond 14C assay. They are more weathered chemically and more dissected than Margaret age deposits, and the degree suggests a pre-last interglaciation age (> 130,000 yr B.P.). The snow line of the ice cap lay at 740 m, and annual temperature was reduced by 7°C. Ice of the earliest Linda Glaciation slightly exceeded that of the Henty Glaciation but had a similar distribution. The glacial deposits are intensely weathered, have reversed magnetization, and overlie a paleosol containing pollen of Tertiary type. An early Pleistocene or Tertiary age is indicated.

Analysis of the geomorphology, geology, and palynology of deposits in the King Valley permits the identification of four glaciations and two interglaciations and has led to a revision of the Pleistocene stratigraphy of the West Coast Range. The oldest late-Cenozoic deposits in the valley appear to predate glaciation, contain extinct pollen types, and are probably of late-Tertiary age. Overlying deposits of the Linda Glaciation show intense chemical weathering and have a reversed detrital remanent magnetization indicating deposition before 730,000 yr B.P. The highly weathered tills are conformably overlain by organic deposits of the Regency Interglaciation which show a transition from montane scrub rainforest to lowland temperate rainforest. Deposits formed during the later Moore Glaciation record advances of the King Glacier and glaciers from the West Coast Range. A pollen-bearing fluvial deposit records an interstade during this glaciation. On the basis of weathering rinds, amino acid dating, and palaeomagnetism the deposits are estimated to have formed between 730,000 and 390,000 yr B.P. The Moore Glaciation deposits are overlain by sediments of the Henty Glaciation which are believed to predate 130,000 yr B.P. These deposits record multiple advances of the King Glacier and the development of a large lake during an interstade. Deposits of the subsequent Pieman Interglaciation consist of organic fine sands and silts that record a lowland scrub rainforest. Deposits of the last (Margaret) glaciation are restricted to small areas in the northern part of the valley. Although the most recent ice advance culminated after 19,000 yr B.P., evidence of older deposits of the Margaret Glaciation suggests that an early last-glaciation ice advance may have occurred. When combined with earlier studies, the recent work in the King Valley has provided one of the more complete records of Pleistocene glaciation in the Southern Hemisphere. Comparison of the deposits with the record of glaciation in southern South America and Westland, New Zealand, suggests some similarities exist between pre-last-glaciation events and indicates that glacial events in Southern Hemisphere middle latitude areas were synchronous during the last glaciation.

Lake Johnston cirque contains some of the best subalpine rainforest in Tasmania. Pollen from the sediments shows Lagarostrobos franklinii, which presently reaches 1040 m, may be a glacial relict. Nothofagus cunninghamii–Nothofagus gunnii subalpine rainforest developed between 9000 and 6000 14C yr B.P., with a maximum at 8700 14C yr B.P. After 6000 14C yr B.P. Nothofagus gunnii became more important, and from 3600 14C yr B.P. sclerophyll and heath components increased. Partial burning of the catchment occurred periodically. Early Holocene climate was warmer and wetter than late Holocene climate. The vegetation and climate changes are similar to those recorded from western South Island New Zealand and Chile. Radiocarbon dates give a sedimentation rate of 0.43 mm/yr. Cores are correlated by magnetic susceptibility. Magnetic ages are assigned by matching with the 14C-dated secular variation master curve for southeastern Australia. Magnetic ages are consistent with the 14C chronology when the former are adjusted by 350 years.

The effects of past glaciation in what is now Australian territory were first recognized on Macquarie Island, probably by David Ramsay, in 1821. The recognition by Darwin in 1836, and reporting by Milligan in 1848 of ice-transported pebbles and boulders in late Palaeozoic marine rocks in Tasmania, showed on the one hand participation in and on the other familiarity with the controversy in Great Britain at that time on the origin of erratics and drift currents. Reports by Clarke (1852), Daintree in 1859, Selwyn (1860), and Gould (1860) of the effects of land ice on Mount Koscuisko (New South Wales), Bacchus Marsh (Victoria), Inman Valley (South Australia), and the Central Highlands (Tasmania), respectively, reflect the increasing recognition in Great Britain of the erosional and depositional effects of glaciers. Daintree, Selwyn, and Gould were all closely connected with A.C. Ramsay, the main British protagonist of the glacial theory at the time, whereas David Ramsay and Milligan were probably influenced by Robert Jameson of Edinburgh.

A Pliocene (2.6–3.5 Ma) age is determined from glacial sediments studied in a 20 m long, 4 m deep trench excavated in Heidemann Valley, Vestfold Hills, East Antarctica. The age determination is based on a combined study of amino acid racemization, diatoms, foraminifera, and magnetic polarity, and supports earlier estimates of the age of the sedimentary section; all are beyond 14C range. Four till units are recognized and documented, and 16 subunits are identified. All are ascribed to deposition during a Late Pliocene glaciation that was probably the last time the entire Vestfold Hills was covered by an enlarged East Antarctic Ice Sheet (EAIS). Evidence for other more recent glacial events of the ‘Vestfold Glaciation’ may have been due to lateral expansion of the Sørsdal Glacier and limited expansion of the icesheet margin during the Last Glacial Maximum rather than a major expansion of the EAIS. The deposit appears to correlate with a marine deposition event recorded in Ocean Drilling Program Site 1166 in Prydz Bay, possibly with the Bardin Bluffs Formation of the Prince Charles Mountains and with part of the time represented in the ANDRILL AND-1B core in the Ross Sea.

The East Antarctica ice sheet advanced onto the continental shelf during the last glaciation but appears to have been thinner (<300 m) than previously hypothesised and probably did not everywhere extend to the edge of the continental shelf. Where the shelf is wide, the ice probably terminated against shallow banks on its outer edge. There may not have been time for the sheet to develop the maximum profile form and thickness predicted by Hollin (1962) and Hughes and others (1981) of about 1000–500 moverthe shelf. Large outlet glaciers occupied deep troughs that conveyed most of the ice towards the edge of the shelf: intervening areas were less intensely glaciated. Much of Prince Charles Mountains and Amery Oasis were not ice-covered: Vestfold, Bunger and Casey oases were glaciated. Vestfold and Bunger oases became ice-free after 10 ka BP under the influence of the Holocene marine transgression, which was complete by about 6 ka BP. During at least the last 5–6 ka these oases have been approximately their present size. Since then the margins of the Antarctic continental ice sheet have maintained almost steady state conditions against the landward edges of the hill masses.

The Bunger Hills were covered by the Antarctic Ice Sheet during the last glaciation. During deglaciation the ice sheet margin collapsed into the marine inlets and the sea entered the oasis before 7.7 ka BP. Raised beaches occur widely below 8.5 m and indicate uplift at 1.4 m ka−1 during the middle and late Holocene. After the coastal inlets were formed, the Edisto Ice Tongue and Apfels Glacier still impinged on land margins in the west of the oasis. Two sets of marginal moraines were formed; the Older Edisto Moraines after 6.2 ka BP and the Younger Edisto Moraines during the last few centuries. The margins of the Antarctic Ice Sheet and Apfels Glacier in the south have maintained their present positions since at least 5.6 ka BP and probably 10 ka BP.

Studies of glacier margins have suggested that form and structure can be used to infer mass balance condition and stability. This paper examines this idea by investigating the form and structure of the Antarctic ice sheet at two coastal oases in East Anarctica. Two principal forms of the margin of the ice sheet in the Vestfold and Bunger hills that are discussed are ice cliffs and gently-sloping ice margins with an inner moraine. The form of the ice margins in both areas is primarily related to the accumulation of drift snow and superimposed ice and not to mass balance condition. It is concluded that the form and structure of ice margins are ambiguous indicators of mass balance condition and stability and that a change in mass budget is probably neither a sufficient nor a necessary condition for a change in the morphology of ice margins. Although we argue that the form and structure of the ice margins tells us little about stability, interpretation of the Holocene history and geomorphology of the oases suggests that the ice margins have been stable at least throughout the Holocene and that this condition of overall stability continues today.

In the Bunger Hills, mapping of glacial drift sheets and examination of striae patterns and other palaeo-ice flow direction indicators show that the largely ice-free region records the imprint of ice sheet expansion(s) during the late Cenozoic. In particular, ice moulded features and striae in southern Bunger Hills suggest formation during at least two episodes of ice sheet expansion, although whether they were formed during separate events or merely different phases of the same expansion of the ice sheet is not able to be discerned at present. The older event relates to thin ice with flow constrained by the topography, whilst the younger event relates to regional expansion of thick ice across the area. Discrimination of the order of emplacement of the cross-cutting striae patterns is possible at a number of sites. Palaeo-ice flow indicators confirm that ice sheet expansion over southern Bunger Hills was purely from the southern and eastern margins, although minor advances of the north-east flowing Edisto Glacier onto coastal areas occurred following retreat of the last extensive ice sheet phase.

This review assesses the circumpolar occurrence of emerged marine macrofossils and sediments from Antarctic coastal areas in relation to Late Quaternary climate changes. Radiocarbon ages of the macrofossils, which are interpreted in view of the complexities of the Antarctic marine radiocarbon reservoir and resolution of this dating technique, show a bimodal distribution. The data indicate that marine species inhabited coastal environments from at least 35 000 to 20 000 yr BP, during Marine Isotope Stage 3 when extensive iceberg calving created a ‘meltwater lid’ over the Southern Ocean. The general absence of these marine species from 20 000 to 8500 yr BP coincides with the subsequent advance of the Antarctic ice sheets during the Last Glacial Maximum. Synchronous re-appearance of the Antarctic marine fossils in emerged beaches around the continent, all of which have Holocene marine-limit elevations an order of magnitude lower than those in the Arctic, reflect minimal isostatic rebound as relative sea-level rise decelerated. Antarctic coastal marine habitat changes around the continent also coincided with increasing sea-ice extent and outlet glacial advances during the mid-Holocene. In view of the diverse environmental changes that occurred around the Earth during this period, it is suggested that Antarctic coastal areas were responding to a mid-Holocene climatic shift associated with the hydrological cycle. This synthesis of Late Quaternary emerged marine deposits demonstrates the application of evaluating circum-Antarctic phenomena from the glacial-terrestrial-marine transition zone.

This overview examines available circum-Antarctic glacial history archives on land, related to developments after the Last Glacial Maximum (LGM). It considers the glacial-stratigraphic and morphologic records and also biostratigraphical information from moss banks, lake sediments and penguin rookeries, with some reference to relevant glacial marine records. It is concluded that Holocene environmental development in Antarctica differed from that in the Northern Hemisphere. The initial deglaciation of the shelf areas surrounding Antarctica took place before 10 000 14C yrs before present(BP), and was controlled by rising global sea level. This was followed by the deglaciation of some presently ice-free inner shelf and land areas between 10 000 and 8000 yr BP. Continued deglaciation occurred gradually between 8000 yr BP and 5000 yr BP. Mid-Holocene glacial readvances are recorded from various sites around Antarctica. There are strong indications of a circum-Antarctic climate warmer than today 4700–2000 yr BP. The best dated records from the Antarctic Peninsula and coastal Victoria Land suggest climatic optimums there from 4000–3000 yr BP and 3600–2600 yr BP, respectively. Thereafter Neoglacial readvances are recorded. Relatively limited glacial expansions in Antarctica during the past few hundred years correlate with the Little Ice Age in the Northern Hemisphere.