9

Traces of the forest

I am inclined to look in the southern, as in the northern hemisphere to a former and warmer period, before the commencement of the last Glacial period, when the Antarctic lands, now covered by ice, supported a highly peculiar and isolated flora.

Charles Darwin, On the Origin of Species, 1859

The ancient forest cover

Site 270 Leg 28; Ross Sea

In the sediments drilled in the Ross Sea there were at least hints of an ancient forest cover; this was what I was hoping to find. At Site 270, closest to the great Ross Ice Shelf, most of the sediments showed evidence of having been deposited by ice. They were hard dark clays carrying the telltale pebbles reflecting their carriage by icebergs—some of the pebbles were up to 10 centimetres long. There was also clear evidence that the sediments had been deposited at sea. Most cores yielded fragments of mollusc shells; there were bivalves and coiled snail-like gastropods, as well as the glassy frustules, the silica-rich cell walls, of diatoms. Most likely the deposit had accumulated from sediment-laden icebergs calving directly into an open, ice-free Ross Sea. The layers close to the bottom of the drill hole were rich in plant debris. There were tiny fragments of leaf and woody tissue, remnants of stem cells and pollen; all of this lovely rubbish suggesting it hadn’t travelled far from a living vegetation source.

I had been classified as a sedimentologist on this cruise because it wasn’t possible to do the pollen analysis on board. Getting a look at the vegetation history from the cores we recovered proved a task needing much patience. To free pollen and spores—palynomorphs—from their enclosing sediments requires complex methods and powerful acids, including hydrofluoric acid, one of the most corrosive known, to dissolve away silica-rich muds and clays. For this reason, sample preparation for palynology was at that time considered too dangerous to be done on board a pitching ship.

So that part of the program had to be carried out ashore—back in the labs of my employer, the Bureau of Mineral Resources in Canberra, now Geoscience Australia. And it was clear that only those cores taken close to the continent were likely to provide the history of the land flora, given that the Antarctic continent itself was the source of the pollen and spores in the seafloor sediments. Cores drilled at a great distance from that landmass were unlikely to yield such valuable debris. They were, however, sometimes rich in dinoflagellates—organic-walled unicellular plankton that are useful in establishing the ages of sedimentary sequences.

We processed 101 core samples to recover their pollen. However, of the six drill sites from which cores were processed, it was only this one, Site 270, that yielded a pollen suite rich enough to shed light on the history of the land flora. Initially, it was hoped that the drill would pass through a significant thickness of sediments beneath those of glacial origin, so that we might see a change in the vegetation from that immediately before the icecap to that growing under dramatically deteriorating climate conditions. But the sediments underlying the glacial ones—greenish-coloured sandstones—were thin and barren of pollen. The green sandstones had their uses, however. They owe their green colour to the mineral glauconite—an iron potassium silicate that typically forms in shallow water. This could be dated geochemically using isotopes of potassium and argon. The date it gave was 26 million years—Oligocene. Elsewhere, in more temperate latitudes, this was a time when grasses became dominant.

But just above the green sandstones, in a grey mudstone with lots of animal burrows and dropped pebbles showing iceberg activity, pollen and spores were abundant, along with a whole lot of degraded plant debris showing that it had been deposited close to an ancient shore.

As always, when confronted with such microscopic imagery, I experienced a sense that lies somewhere between disbelief and privilege. Did these fragments, these objects invisible to the naked eye, sometimes degraded and ragged, but occasionally showing the precise and beautiful structure of pollen, or the cells of a piece of leaf tissue, really represent a forest whose age and nature we can barely comprehend? Was this as close as I could get to this forest floor? The privilege lies in having the equipment, the opportunity and the knowledge to be afforded just a glimpse of an ancient, pre-human landscape.

Above the green sandstones, all the sediments contained pollen. Grains of the southern beech Nothofagus were there in abundance. This tree grows today across a broad sweep of continents, mostly in high southern latitudes such as Chile, parts of Argentina, eastern and southern Australia (particularly Tasmania) and New Zealand, but it sneaks too into the cooler parts of the tropics in New Caledonia and the highlands of New Guinea. Trees of Nothofagus were once included in the Fagaceae—the northern hemisphere family of beeches and oaks and chestnuts —but botanists now place them within their own family, Nothofagaceae, sometimes translated as the ‘bastard beeches’.

Figure 9.1. Nothofagus gunnii in Tasmania in autumn foliage.

Source: Wikipedia Commons.

The genus Nothofagus is further broken up into groups that show up in the pollen record and are probably significant in terms of past climates. For example, the group Brassospora that produces pollen of the brassi type seems today to grow mostly in the tropical or subtropical parts of the range of Nothofagus. Other types seem to be linked to the cooler parts of the modern range, with some living species being deciduous, such as Tasmania’s deciduous beech, Nothofagus gunnii (Figure 9.1). The pollen at Site 270 was mostly of the cooler climate type of Nothofagus; pollen linked with subtropical trees was less common. The groups of Nothofagus were originally called sub-genera, but it has recently been argued that, on the basis of DNA, they should be elevated to the level of genus. Should this be accepted, then Nothofagus gunnii would now be Fuscospora gunnii. Here for convenience I continue to use the well-known genus name Nothofagus.

Pollen of Proteaceae was present in the Ross Sea borehole but it was sparse, consisting of only a few simple forms. In contrast, pollen floras of similar age in the Australian fossil record show abundance and diversity of this group. There was also the small pollen of the family Myrtaceae—but this could only be identified at family level, so no claims could be made that Eucalyptus was present.

The southern conifers—mostly in the family Podocarpaceae, or podocarps—were also common in the pollen suite; some looked like that of the small shrubby strawberry pine (Microcachrys) living now in Tasmania, some resembled pollen of the Huon pine (Lagarostrobus) from the same area, and some were like the pollen of the celery-top pine (Phyllocladus). There were a few fern spores, possibly like some of the modern tree ferns—and rare mosses.

What would this vegetation have looked like? It might have been a forest with a tree cover of southern beech and podocarps, and an understorey of ferns and mosses. Alternatively, and perhaps more likely, given that the vegetation grew alongside nearby glaciers, it may well have consisted of patches of stunted forest—more tundra-like in aspect—but without the low-growing herbs and grasses that distinguish modern tundra.

As with most pollen found in offshore Antarctic sediments, there is always the possibility that the waxing and waning of the ice sheets has stirred up and mixed the pollen, so that grains of different ages might be tumbled together, making the interpretation of the original vegetation very tricky. This might have been the case at Site 270. Since the time of that study, new approaches have been developed to clearly separate pollen of different ages. These use the way in which pollen grains fluoresce under ultraviolet light. The amount of fluorescence changes with time: older grains lose this capacity to glow; the degree of red fluorescence colour, measured quantitatively, can show whether or not particular pollens are in place or recycled.

Putting together a vegetation history: What can the pollen story tell us now?

Since the drilling of Site 270 on Leg 28, further drilling and sampling programs have yielded much more information on the vegetation history of Antarctica. With more drill sites off the eastern margin of the continent, and a number of independent sites drilled within the Ross Sea, sometimes from sea-ice, plus sampling from field parties—especially from sites close to the Transantarctic Mountains—a picture is now beginning to emerge of a land vegetation shifting from a subtropical forest, through a series of tundra types, then to dwarfed depauperate communities, before the eventual extinction of all higher plants from the continent.

In this section I have given, in chronological sequence, the story of Antarctic vegetation in the Cenozoic, as based on the pollen record, from the oldest assemblages, in the Eocene, to the very youngest, somewhere in the Miocene or Pliocene.

It was the far side of the continent—in Prydz Bay, a deep indentation on the coastline of East Antarctica—that was to give a clearer picture of at least part of the vegetation story. Further cruises of the Ocean Drilling Program were the source of this new information. In the summer of 1987/88 the program again ventured into Antarctic waters, this time with the drilling ship the Joides Resolution—the successor to the Glomar Challenger.

Cruise Leg 119 was drilled in Prydz Bay close to the continent. The wide embayment of Prydz Bay is fed by the Lambert Glacier, feeding in through the Amery Ice Shelf, an ice stream system that drains around one-fifth of Antarctica’s ice. Boreholes within the bay could be expected to provide a long-term history of glaciation on the continent. While these sites, at 68^oS, were close to the continent’s edge, they were further north than the Glomar Challenger’s drill sites in the Ross Sea.

The sediments drilled in Prydz Bay clearly showed that glaciation had begun there even earlier than what we were able to glean from our Leg 28 data—a figure of possibly 36–40 million years was suggested to be the time when ice first developed at sea level. Several major advances of the ice sheet beyond its present limits followed. I was asked to look at the pollen record of four of the boreholes in Prydz Bay. The results were disappointing, as most cores showed several generations of pollen and spores—they showed what was becoming the now common story of the fossil debris being recycled, jumbled by glacial action.

Before the cooling; tropical forests of the Wilkes Land coast (Early Eocene; 48–55 million years)

Pollen from sites in Prydz Bay suggested that there was ice at sea level in the Late Eocene. Since those sites were drilled, the Integrated Ocean Drilling Program (IODP) drilled cores from further east, off the coast of Wilkes Land in East Antarctica. One site, drilled during Expedition 318 in the summer of 2010, found pollen-rich sediments of Early Eocene age—estimated to lie somewhere between 48 and 55 million years old.

This was a vegetation growing in a ‘greenhouse world’! In the time of this flora the differences in temperatures between the equator and the poles would have been much less than they are now, and the levels of atmospheric carbon dioxide much higher—something in the vicinity of 1,000 parts per million.

Forests of subtropical to tropical aspect were clearly growing there in the Early Eocene. A rich pollen and spore flora from a site labelled U1356 shows this; geochemical data from fossil soils at the same site confirms it. At that time, the latitude of the site was about the same as it is now, so that this vegetation would have endured around 50 days of polar darkness each year. Yet there was a great diversity of flowering plants—today these are part of tropical or subtropical forests in Australia, New Guinea and New Caledonia. The structure of the vegetation in the Early Eocene probably had a canopy of the kind of trees that occur today in tropical settings; below that an understorey of ferns flourished.

Significantly, there was the pollen of palms—palm trees grow now only in tropical and temperate parts of the world. They will only tolerate conditions where temperatures during the coldest month don’t drop below 5^oC. Another group typical of warmth-loving tropical forests belongs to the Bombacoideae, a subfamily that includes the baobobs or bottle trees, the silk cotton trees (Bombax) and the kapok trees (Ceiba). While these are rare in the pollen assemblages of Wilkes Land, they are known to be insect-pollinated and their pollen usually falls within 100 metres of the tree, so finding them in the samples suggested they were an important part of the forest vegetation. These are the southernmost records of these tropical trees. Winter temperatures must have been substantially above freezing for this part of the Antarctic coast. Mean annual temperatures then might have been as high as 16^oC for this Early Eocene part of the record.

But pollen of more temperate rainforests, including the beeches and the podocarps, is present too, mixed in with the tropical pollen in the lower part of the section. Perhaps the parent trees of these grew on higher parts of the landscape, or further inland. These temperate rainforest trees came to dominate in the Middle Eocene, when the more tropical rainforest became extinct in the area. This showed that coastal regions of Wilkes Land cooled strongly then, and mean annual temperatures dropped to as low as 9^oC. These cooler rainforests were likely to have been the parent vegetation that gave rise to the alpine heaths or scrubby rainforests of latest Eocene age that we encountered far to the west in Prydz Bay.

Striking it lucky: Cooler forests in the Late Eocene of Prydz Bay (34 million years)

In the millennial summer, January to March of the year 2000, the Antarctic sea floor was drilled again in Prydz Bay, during Leg 188 of the ongoing Ocean Drilling Program. One of the boreholes, drilled in the centre of this great indentation in the coast, gave an unparalleled record of this late stage Antarctic vegetation—vegetation that immediately preceded the development of a major icecap. This was the now legendary Site 1166. Cores taken there came from sedimentary strata rich in organic matter—mostly broken plant debris, including pieces of wood and well-preserved pollen and spores. But mixed with this were also dropstones—pebbles that signalled ice in the vicinity. The vegetation reflected there may have been growing on delta habitats or on the sides of channels within the bay. But wherever it grew, it clearly coexisted with ice at sea level. The age of this sequence, based on the time-ranges of other microfossils, including dinoflagellates, lies somewhere close to 34 million years—in the youngest bit of the Eocene.

In compiling the details of the parent plants reflected in this pollen assemblage, I was joined by Dr Mike Macphail of the Department of Archaeology and Natural History at The Australian National University. His knowledge of living floras was invaluable in reconstructing the ancient vegetation.

It was a diverse flora that grew along with this first ice. We documented more than 80 species of flowering plants, 20 gymnosperms (including conifers and cycads) and some 25 cryptogams—the ferns, mosses, liverworts and algae. Pollen counts showed that the southern beech, with five different species, and the southern conifers or podocarps, with six different species, again dominated. Other conifers included araucarias—a family now predominantly of southern hemisphere distribution. Some kinds of pollen in that group possibly represent the Wollemi pine. Pollen akin to that of some modern cypresses was there too. Did these represent trees or had they been reduced to mere shrubs in this hostile environment? We could only speculate.

Among the flowering plants, pollen of Proteaceae was most abundant. Other plant families included rare Myrtaceae (but not Eucalyptus) and pollen of Droseraceae, the family characterised now by the carnivorous sundews and the Venus fly traps. Its presence in Antarctica in the Eocene suggests that insects must have been there supplying nitrogen where soils were deficient in nutrients. There are likely to have been ericas, too, and casuarinas, and members of the family Caryophyllaceae—the carnations and pinks—of which a single genus, Colobanthus, grows now in low-latitude parts of Antarctica. There was also pollen of sedges, of reeds, rushes and possible lilies.

What were the plants that made up this late stage vegetation? And what might their structure have been, surviving under the toughest of conditions, enduring not only near-freezing temperatures but also long periods of winter darkness?

Collectively, all these plants mean we might be looking at something akin to a modern alpine heath in Tasmania. At first we called it a ‘scrubby rainforest’ but, on a second examination, ‘alpine heath’ seemed more appropriate (Truswell and Macphail 2009).

Figure 9.2. Eocene pollen from sediments in Prydz Bay: a. pollen of Araucariaceae; b, c. Nothofagaceae—pollen of the Nothofagus fusca type; d. Proteaceous pollen; f, g. biwinged pollen of the conifer family Podocarpaceae.

Source: Elizabeth Truswell and Mike Macphail.

Species of rainforest trees grow in these modern alpine heaths but they are reduced in stature and are more common below the treeline. Mike Macphail surmised that the Prydz Bay vegetation might have been a mosaic of Krumholtz trees—the name given to stunted and deformed trees growing in subalpine habitats in the northern hemisphere—in combination with lower growing shrubs and herbs. It is possible to visualise vegetation something like that referred to in the northern hemisphere as ‘taiga’, a mix between conifer forests and the Arctic tundra. Some plants might have been deciduous, although a seasonal leaf fall is not as common today among southern forests as it is in those of northern climes.

Whatever the plant community, it had to survive dark conditions in winter, probably poor soils and high rainfall. The dark winters may not have proved as much of a barrier to plant growth as one might expect; laboratory experiments have shown that some species growing now in cool rainforests can survive long periods of darkness. If the vegetation was relatively open it has been estimated that the light energy available at ground level on a yearly basis would not differ significantly from that received in Earth’s more temperate regions. The high rainfall suggested by the nearest living relatives of many of the plants identified should not have been a problem; many grow today under high rainfall conditions, so that rainfall in the range of 1,200 to 1,500 mm per year seems possible. The temperatures under which this vegetation grew are more difficult to estimate, but we have suggested that mean annual temperatures of less than 12^oC were likely.

Persistence; the vegetation struggles on

The Ocean Drilling Program (ODP) borehole 1166 in Prydz Bay clearly shows that plants were growing there alongside ice at sea level, close in time to the Eocene–Oligocene boundary, some 34 million years ago. This time is now recognised as close to the point when Earth left its warmer greenhouse state and entered the present ‘icehouse’, distinguished by an icecap on Antarctica—and perhaps ice elsewhere. This change was one of the most dramatic in Earth’s climate history.

The history of the icecap, and what caused it to form in the first place, remains problematic and subject to a great deal of debate. As discussed earlier, the transition from greenhouse to icehouse was formerly considered to coincide with the beginning of the Antarctic Circumpolar Current that now circles the continent, causing Antarctica to be thermally isolated from warm waters flowing from the north. As always in geology, the picture has become increasingly complicated. As outlined more fully in Chapter 7, the glaciation of Antarctica is no longer solely linked to the initiation of the current. As we saw previously, other factors probably came into play to cause the formation of ice at sea level—factors relating to the decline of atmospheric CO₂ levels, to albedo effects of ice, even to vegetation cover—all may have been a part of this dramatic change.

What happened to the distinctive flora we discovered at Prydz Bay? How long did it persist under increasingly adverse conditions? Were there shrubs and low trees that survived the expansion of the ice sheet of East Antarctica to sea level? Unfortunately the Prydz Bay drill site can’t tell us that because the top of the pollen-bearing section there has been planed off by the development of the Amery Ice Shelf. To understand the fate of this rather impoverished Late Eocene vegetation we must return to sites in the Ross Sea region. In The Vegetation of Antarctica through Geological Time (2012) authors David Cantrill and Imogen Poole have given a detailed and systematic account of the vegetation history after the Eocene. Their final descriptive chapter is aptly called ‘After the Heat: Late Eocene to Pliocene Climatic Cooling and Modification of the Antarctic Vegetation’.

Vegetation in the Ross Sea region—Eocene, Oligocene and Miocene

Palynological studies carried out by Lucy Cranwell and co-workers in the 1960s were the among the first to be undertaken in Antarctica (Cranwell et al. 1960). These were on erratic boulders in moraines—accumulations of rocks at the edges or ends of glaciers—in the area of McMurdo Sound in the southwest of the Ross Sea. Some of these erratics proved to be of Eocene age, with pollen assemblages that reflected a rich flora. Nothofagus was dominant, along with other angiosperms (Proteaceae, Casuarinaceae, Ericaceae, Restionaceae (rushes) and, more rarely, Myrtaceae and lilies). There were podocarps, too, and Araucaria, Ginkgo, cycads and ferns. This floristically rich vegetation has been compared to forests in the Araucaria region of Chile, at 38^oS. They were a little older (perhaps mid- to late Eocene) than the Prydz Bay floras of the east coast of Antarctica, and perhaps more diverse.

From the summary by Cantrill and Poole, it is evident that ice-free areas were present in the Ross Sea region through much of the Oligocene and Miocene (i.e. roughly from 30 to 10 million years ago), and that some of the shrubs, or shrubby trees, that had distinguished the Eocene persisted for variable lengths of time in this broad area. Some few, perhaps half a dozen, may have survived in ice-free, possibly sheltered areas within the Transantarctic Mountains as recently as the Pliocene, some 3–5 million years ago.

Pollen assemblages of Oligocene age have been found from the series of boreholes drilled to explore the history of the Ross Ice Shelf, and detailed in the preceding chapter. Cores from the multinational Antarctic Geological Drilling (ANDRILL) boreholes, the series of boreholes at Cape Roberts and an earlier borehole, Cenozoic Investigation in the Western Ross Sea (CIROS-1), all yielded a dominance of Nothofagus pollen throughout the Oligocene and into the Early to Middle Miocene—to about 15 million years ago. It is clear from the nature of the remains that forests of southern beech were at times growing locally. Some of the pollen was clinging together in clumps, just as if it had been released from whole anthers—these must have been shed from nearby trees and not recycled from older deposits. A leaf was found in the Early Miocene part of the section in the CIROS-1 borehole, supporting this claim; it bore a pattern of veins similar to that shown by the living deciduous species Nothofagus gunnii from the highlands of Tasmania. Conifers—Podocarpus—were there in the pollen suite, too, and a minor component of flowering plants, including some Proteaceae, grasses and daisy-like forms. These may have grown in rocky pockets within a vegetation whose closest modern equivalent could have been the Magellanic subpolar forests of Patagonia in southern South America. The picture was that of a tundra with herbs and mosses and, locally, clumps of shrubby trees. All would have grown under high rainfall, with freezing temperatures in winter but with summer temperatures perhaps not falling below 5^oC.

Other boreholes, including those at Cape Roberts, produced similar floras, perhaps of lower diversity. The Cape Roberts boreholes showed marked fluctuations in the abundance of pollen: samples with more pollen grains per gram of sediment have been interpreted as warmer phases with more abundant and dense vegetation; those with fewer grains per gram suggest that the total mass of vegetation was less. These fluctuations have tentatively been tied to the periodicity of Earth’s orbit around the sun, with a maximum in the eccentricity of the orbit that occurs every 400,000 years. This is usually considered to reflect warmer conditions.

The picture evoked by all of this pollen data is one where relatively tall forests of Nothofagus gave way to a more shrubby and open forest with decreased diversity. Again, today’s Magellanic forests, or the temperate rainforests of Tierra del Fuego, growing next to coastal glaciers might be appropriate, though not precise analogies.

Clearly, this flora no longer exists on the Antarctic continent. The pathway to its later extinction is far from clear, but appears to have had some bends and twists. After the Oligocene, there was a phase in which Antarctic climate warmed again, around 15–17 million years ago, in the middle of the Miocene. The warming is reflected in the fossil flora of the Ross Sea region; the drill site data from the ANDRILL boreholes shows a flora in which the shrubby forms became more tree-like—beech and podocarps proliferated, but other flowering plants, perhaps as groundcovers, included members of the pink family, trigger plants, sundews and ericas, grasses, reeds and rushes, plus an abundance of mosses and liverworts. At this time summer temperatures are suggested to have reached, or possibly exceeded, 10^oC. Abrupt cooling and a moss-rich tundra vegetation, wherein woody plants were again represented by shrubs, followed this warm phase.

The warming in the Middle Miocene shows up too in the Dry Valleys—the snow free valleys in the Transantarctic Mountains to the west of McMurdo Sound. Glacial deposits there have formed dams that contain small lakes. The sediments are interbedded with volcanic ash layers. These have been confidently dated, by using isotopes of argon present in volcanic ash, as 14.7 million years old. They yielded fossils of mosses—exquisitely preserved and looking as though they had been instantaneously freeze dried, rather than fossilised; diatoms were there, as were algae and shrimps. The terrestrial flora reflected by pollen suggests domination by a single species of deciduous beech—this may have been the only tree-like plant. There is no conifer pollen, so the vegetation may have been sparse. But there were also members of the pink family and again mosses and liverworts. Insect remains include parts of a small beetle. Overall, one could describe the vegetation as ‘tundra-like’ and growing under climates much warmer and wetter than those of today.

For reasons of space and complexity I have not discussed the fossil floras of the Antarctic Peninsula, which lies at a lower latitude than the rest of Antarctica. Some of the islands of the peninsula, which were at latitudes of around 62^oS in the Eocene—sites from Seymour and King George islands for instance—have yielded rich leaf fossils as well as pollen. Diverse Nothofagus and conifer forests described from these have been compared to forest and fern bush vegetation growing now on southern ocean islands. But the Oligocene and Miocene sequences from the islands show a reduction in diversity similar to that on mainland Antarctica.

Very near the pole; a last hurrah?

The most controversial of all the plant deposits in the Antarctic record is no doubt the flora found in sediments of the Sirius Group, known from some 40 localities scattered along the inland flanks of in the Transantarctic Mountains. They are possibly Pliocene—around 3 million years in age; but could be as old as Miocene—17 million years old. The younger dating was based on the presence of marine diatoms, but one argument suggests that these may have been blown in at a later date. The Meyer Desert Formation, the uppermost part of the Sirius Group, occurs as packages of sediment sitting atop a glacially scoured erosion surface. The deposits extend as far south as latitude 86^oS. Thus they lie within 500 kilometres of the present South Pole. But the implications of plant growth at this time and this latitude are dramatic, implying that the ice sheet was unstable and responded to a warming phase by retreating from its present extent. However, the geography and the dynamics of ice sheet retreat remain controversial; a favoured possibility is that these deposits, now at levels high in the Transantarctic Mountains, may have been deposited on the margins of a fjord close to sea level and later uplifted. The 3 million year dates, if confirmed, would imply ice sheet instability, which may also be related to the expansion of ice sheets in the northern hemisphere.

The fossil assemblages from these deposits in the Transantarctic Mountains are unexpectedly diverse, as shown by the fossil assemblages from Oliver Bluffs. The flora and fauna was a highly specialised one, and relatively well-preserved, with leaves and wood of Nothofagus, and flowers, fruit and seeds of other vascular plants, as well as five species of moss. Among the vascular plants some are represented by pollen; others by more substantial remains—grasses, sedges, buttercups, mares-tails, chenopods (the family that includes the saltbushes) and/or possible Myrtaceae. Bacterial mats and algae contributed to the ecosystem. Animal life includes weevils, snails and clams. No macrofossil conifers were evident, although rare pollen grains of Podocarpaceae were present.

The most dramatic plant fossils are again the southern beeches. Here they are spectacular because they assumed growth habits similar to those of prostrate species known from the extremes of southern South America. Trees are dwarfed and prostrate—perhaps ankle-high—as shown by woody fossils. They had diameters rarely greater than a centimetre; they appear to have grown slowly, and become contorted, even entwined around cobbles. Unsurprisingly, they were deciduous, with patterns in the leaf veins again similar to the deciduous Tasmanian Nothofagus gunnii although the leaves are not identical. And they also accumulated in dense deposits that suggest an autumn leaf fall. This tundra vegetation may have been widespread across some hundreds of kilometres along the Transantarctic Mountains.

The fossil leaves from the Sirius Group have been described as a new species (see Figure 9.3), Nothofagus beardmorensis, the species name referencing the Beardmore Glacier, some 200 kilometres inland from the fossil site. Although formal description has focused on the leaves, the woody fragments (Figure 9.3; upper image) and the abundant pollen in the sediments are considered likely to have been part of the same species.

Figure 9.3. Wood fragments (above) and leaf impressions of Nothofagus beardmorensis (below) from the Sirius Group.

Source: Courtesy of Robert S. Hill, University of Adelaide.

Recent geochemical studies have provided new evidence on the nature of the flora and the conditions under which it might have survived. Biomarkers, in this case tetraether lipids, extracted from sediments at the Oliver Bluffs site have served both as palaeothermometers and as indicators of the plant groups present. Significantly, some of the compounds analysed suggest that conifers, probably podocarps, were indeed present in the vicinity, corroborating the evidence from rare pollen grains. The lipids further suggest that summer temperatures may have been in the vicinity of 5^oC.

These findings are in line with summer temperatures deduced earlier from the known tolerances of the fossil plants. Mean annual temperatures are estimated to have been minus 8^o to minus 12^oC, with summer months as warm as perhaps 4–5^oC. Temperatures in the rest of the year could have dropped as low as minus 22^oC. The growing season may have been as short as 12 weeks. Strong winds and frosts through the growing season would have increased the severity of that short season. During such a period, soils would have been able to form, aided by colonising cushions of moss that would have trapped organic matter.

There is just one other hint that Nothofagus-dominated forests might have persisted into the Pliocene. That comes from Site 274, the last drill hole of the Leg 28 cruise. It was not examined for its pollen until 1996, when it was thought that information from that site might be useful in solving the riddle of Antarctica’s last vegetation. An age of Pliocene for part of the core was established using diatoms. That cored interval also yielded pollen of Nothofagus, notably of five different species types. It was thought to be in place, rather than recycled. If this should be so, it would suggest Nothofagus forest was still present on the adjacent coast of Antarctica as recently as 3 million years ago.

In summary; the road to extinction

The vegetation history of the Antarctic continent—sketched briefly because of the few sites available—shows the presence of a diverse subtropical forest in the Early to possibly Middle Eocene (48–55 million years) of the Wilkes Land coast. This may be equivalent to that shown by some erratic boulders near McMurdo in the Ross Sea region. By the end of the Eocene (around 34 million years), there was glacial ice at sea level and this forest was reduced to an alpine heath, or a very scrubby rainforest, a situation demonstrated at Prydz Bay.

The Oligocene record, best known from boreholes in the Ross Sea, shows forests giving way to tundra, with clumps of shrubby trees—likened to the living subpolar forests of Patagonia. A period of warming in the Middle Miocene (14–15 million years) locally increased the stature of the tree cover in the Ross Sea area, but in the Transantarctic Mountains the record was one of a sparser, more tundra-like vegetation. The youngest available record—of contested age—showed both diversity and stature of vegetation being reduced further, with ‘trees’ knee-high and contorted in response to decreasing temperatures and seasonal availability of water. The influence of these factors was such that not even tundra could long survive.

This record from the Sirius Group in the Transantarctic Mountains is, as far as we can determine, the last land vegetation, the last vegetation of higher plants, known from East Antarctica. But the age of the record is controversial, and further investigations are needed before we can know just when plant life and its associated biota became extinct from most of the continent. It might have followed soon after the warming phase of the Middle Miocene, or it might have been as late as the Pliocene, a mere 3 million years ago, if the pollen evidence from Leg 28 Site 274 is substantiated. This event, the elimination of all higher plants from the continent of Antarctica, should it be more confidently dated, would represent one of the greatest of all known extinctions.

I have summarised the story of Antarctica’s land vegetation in Figure 9.4. This record, from Eocene and younger sequences is set there against a commonly quoted curve that has come to bear the name of Jim Zachos, the first author of the paper in which it was first published in the year 2001. An updated version appeared in 2008. The curve shows temperature changes on a global scale, using oxygen isotope data based on foraminifera from a large number of deep sea drilling sites. This data has been used to calculate both past temperatures and the volume of ice. The dating is based on a time scale built up from magnetic reversals and the fossil record. The palaeotemperatures shown beneath the curve are given only for the ice-free intervals, before the initiation of Antarctic glaciation.

In relation to what we now know of the Antarctic vegetation story, the tropical or subtropical vegetation known from Wilkes Land correlates with the warmest temperatures shown on the curve—those of the Early to Middle Eocene; the rapid jump into a cooling world, with the presence of ice at sea level is clear at the Eocene/Oligocene boundary with the alpine heath or scrub. In the Miocene a warming is visible around Middle Miocene, with a prolonged tundra-like vegetation dominated by Nothofagus and the podocarps. Then follows more frigid temperatures through the Late Miocene/Pliocene, with patchy groundcovers of the same taxa. The last evidence, currently, is that of the cushion forms and dwarfed distorted trees as in the Sirius Formation. Beyond, there is as yet no evidence for the presence of a higher vegetation.

Figure 9.4. Zachos Curve of global temperatures set against a summary of the vegetation record from East Antarctica. The curve is based on oxygen isotope data from foraminifera drawn from deep sea drilling sites.

Note: Ma refers to millions of years ago.

Source: Drafting by Clive Hilliker.

Old pollen on a modern sea floor

Before I took up the position at Florida State I had become aware that today’s sea floor around Antarctica was virtually littered with pollen and spores from ancient plants that had grown on the continent. I had found these in surface sediments dredged up from the sea floor by earlier expeditions. Some examples from the Australian Antarctic Expedition under Sir Douglas Mawson in 1914–15 are described in Chapter 7.

The surface of the Ross Sea floor is in places also dense with recycled pollen. They are the result of pollen-rich sedimentary rocks in the source area of glaciers being picked up by the action of moving ice, then melting out when the glaciers reach the sea and come into contact with the sea floor at the grounding line. Counts of the abundance of recycled pollen, the number per gram of sediment, in muds dredged from the sea bottom show clear patterns. These are probably related to their transport in rapidly moving ice streams within the Ross Ice Shelf. Through the work of glaciologists, we know something of the nature of these ice streams—they are corridors through the great ice shelves where ice is flowing faster than that of its surrounds. The ice streams fluctuate in time and space. Understanding the nature and history of their flow is important in assessing how well the ice shelves will buttress the melting of continental icecaps under conditions of future warming.

I was fortunate to work with the English glaciologist Dr David Drewry of the Scott Polar Research Institute in Cambridge, who described the ice streams within the Ross Ice Shelf, so I could relate those to the density of pollen on the sea floor. Counts of the pollen showed it to be abundant in the eastern sector of the Ross Sea and distinctly more sparse in the west near the Transantarctic Mountains. Within the eastern sector there was a particular concentration close to where Ice Streams D and E—now called the Bindschadler and MacAyeal ice streams—debouch into the southeastern Ross Sea. This suggests that the source of these sea floor deposits lies in strata in sedimentary basins of Cenozoic age hidden beneath the West Antarctic Ice Sheet in Marie Byrd Land.

While the surface of the sea floor in the Ross Sea and areas offshore from East Antarctica carries a load of pollen debris, this is jumbled and recycled material. Nevertheless, it is useful information. This essentially recycled pollen is serving two aims. First, it is providing a checklist of plant taxa that once grew on Antarctica; this can hint at plant migrations in the past. Second, it is a useful geological tool to suggest where sedimentary rocks of particular ages—or indeed sedimentary basins—might lie beneath the present ice cover. Such information is increasingly valuable in predicting possible future patterns of icecap instability and melting.