Table Of Content

Tertiary Climatic Evolution and the Development of Aridity in Australia Helene A. Martin School ofBiological Science, University ofNew South Wales, Sydney NSW 2052 Martin,H.A. (1998).TertiaryclimaticevolutionandthedevelopmentofaridityinAustralia. ProceedingsoftheLinneanSocietyofNewSouth Wales119, 115-136 The climatic record is deduced from palaeobotany and some other relevant studies. TherecordfromsoutheasternAustraliais the mostcomprehensive andispresentedforcom- parisonwithotherpartsofAustraliawheretherearefewstudies.Inthesoutheast,theprecipi- tationwaswellabovethecriticallevelforrainforest(1500mm)duringthePalaeogene.Inthe mid-late Miocene, precipitation declined to less than 1500 mm and in the late Pliocene- Pleistocene,therewasafurtherdecreasetoabout500-800mm. InthenortheastofAustralia, the precipitation remained well above the critical level forrainforestthroughout the Tertiary andmostoftheQuaternary, hence thisregion is arefugeformany rainforesttaxa. Whenthe northwest of Australia is compared with the southeast, the climate ofthe former was drier thanthelatterthroughoutthelateTertiaryandQuaternary. Grasses were rare until the late Miocene when they show a steady increase which is maintained through the remainderofthe Cainozoic. Grasslands developed first in the north- westand,presumably, CentralAustralia, butdevelopmentwas laterin the southeast. Atrend towards aridity startedinthe midMiocene andcontinuedthroughthe lateTertiary. A degree ofaridity was reached about the late Pliocene-early Pleistocene, but aridity has intensified duringtheQuaternary,especiallyinthelatterpart. Manuscriptreceived20May 1997,acceptedforpublication 19November 1997. KEYWORDS:Aridification,Australia,grasslands,palaeoclimate,Tertiary,vegetationhistory. INTRODUCTION Weather and climate have a daily fascination. The extremes cause discomfort that lead some to claim 'the climate is changing'. While it is difficult to distinguish short term variation from a trend or change on the basis ofthese small scale observations, itis clear that climate has changed, and has changed drastically in the past: the Australian deserts once supported luxuriant vegetation. This study reconstructs the Tertiary climate from the late Eocene, some 40 million years ago, when Antarctica was mostly ice free and Australia was almost entirely forest- ed, and covers the crucial development of aridity. By the end of the Tertiary, some 2-3 million years ago, the climate resembled the modern status, but aridity has intensified, especially in the last halfmillion years. The history ofthe vegetation is amajor source of evidence of climatic change, but there are other studies which complement this history and add invaluable insights into past environments. These past changes suggest that the climate is likely to change in the future, and it could change dramatically from entirely natural causes. The history of the vegetation is constructed from fossil evidence, mainly from palynology, and to a lesserextent, from macrofossils. Each source has both strengths and limitations. Pollen may be extracted from bore samples, hence palynology accesses sedi- ment which may be hundreds ofmetres deep and the record from this source is compre- hensive. Fossil leaves, the major source ofmacrofossil data, however, must be recovered from surface outcrops, although cuticles may be found in bore samples. Both pollen and Proc.Linn.Soc.n.s.w., 119. 1998 6 TERTIARYCLIMATEINAUSTRALIA 11 leaves may show affinities with living taxa, but identification with a modern species is not always possible. Affinities are usually made with genera orfamilies, but afossil pop- ulation may not coincide with any living taxon, even when natural affinities are evident. Dust whipped up by wind may be blown out to sea and deposited in marine sediments. The dust fluxes in deep seacores record the intensity ofdeflation on land (Glasby 1971; McTainsh 1989; Hesse 1994). Phytoliths, the silica bodies produced by grasses and many other plants, may form part of the dust flux and are another source of evidence. Palaeotemperatures deduced from oxygen isotope analyses of deep sea cores (Shackleton and Kennett 1975; Savin et al. 1975), corrected for continental ice volume where necessary (Feary et al. 1991; Isern et al. 1996), indicate surface sea temperatures and the trends may be extrapolated to the land. These records are independent of palaeobotany and hence are invaluable forconfirmation ofthe record from palaeobotany, as well as filling in gaps in the record. Reconstruction ofthe vegetation is achieved from the floristics ofthe palynofloras. The ecological tolerances and climatic limitations ofpresent day taxa and vegetation are applied to the fossil assemblage to deduce past climates. Leaf physiognomic characters, such as dimensions, the nature ofthe margin, drip tips etc. are a direct expression ofcli- mate, irrespective oftaxonomic affinities. The nature ofthe sediments also contains acli- matic signal that is used as supplementary evidence. Numerous experimental studies show that most ofthe pollen in an assemblage at any A one site has beenproduced locally, within aradius ofhalfakilometre orless. little pollen mayhave beentransported in from long distances. (Forreviews ofthe numerous studies on this subject, see Birks and Birks 1980, andMartin 1993.)Thepollenassemblage is areflec- tion ofthe dominanttype ofvegetation and small, isolated patches ofadifferentkind, such as may be found in sheltered gullies and gorges, may be palynologically invisible (Ladd 1979). Pollen in ocean cores, however, has been transported from land either by wind or runofffromrivers and studies ofmarine surface sediments show thatthe pollen inthemis a general reflection of the regional vegetation on land (Mudie 1982; Turon 1984; Prell and vanCampo 1986; Heusser 1988; Mudie andMcCarthy 1994). All kinds of plant parts require the anaerobic conditions of lakes, swamps, bogs etc. for preservation. The vegetation growing around these sites, which occupy the low- est part of the topography, are best represented in the fossil record. Plants growing on sites distant from the sedimentary basin, such as the hilltops or steep slopes usually have little chance ofbeing fossilised. The vegetation units identifiable in the fossil record are usually fairly broad and general. The units of vegetation used in this study are those in common usage and are defined thus: Rainforests Rainforests or closed forests (Specht 1970) are usually structurally and floristically complex. They are found in the better watered environments. Eucalyptus is not a normal part of rainforests. There are many different types of rainforest in tropical, subtropical and temperate regions. In Australia, tropical rainforests are best developed north of 21° latitude, although patches may extend further south. Subtropical rainforests are most common between 21° and 35° south and temperate rainforests are the dominant type south of35° (Webb 1959), although small patches ofthe latter may be found as far north as 28°. Nothofagus is found in temperate rainforests which are especially well developed in Tasmania, New Zealand and Chile, extending to about 50-55°S (Riley and Young 1972). Rainforests may be divided into 'wet', with Nothofagus, and 'dry', usually with various gymnosperms (Kershaw et al. 1994). In drier locations, some scrubs and vine thickets are closely related floristically to rainforests and hence are included in this clas- sification (Webb 1959; Webb and Tracey 1981). Proc.Linn.Soc.n.s.w., 119. i998 H.A.MARTIN 117 Sclerophyll forests or open forests Sclerophyll or open forests (Specht 1970) are most commonly dominated by Eucalyptus and/or Casuarina/Allocasuarina. Wet sclerophyll forests have rainforest or other mesic taxa in the understorey layers, and if left unburnt and undisturbed, may revert to rainforest (Ashton 1981). Dry sclerophyll forests do not have rainforest taxa. The understorey in these forests contains sclerophyllous shrubs (Gill 1981). Fire is an integral part ofthe environment ofsclerophyll vegetation (Ashton 1981; Gill 1981). Open vegetation Inforests, the tree canopy forms an almostcontinuous layer whereas in woodlands, the trees are well spaced. Ifthere are few trees, the vegetation is more open and there is a well developed layerofshrubs, grasses and/orherbs. With adense tree cover, insufficient light filters through the trees to support the ground layer (Specht 1970). In this study, open vegetation refers mainly to shrubland, grassland, and/orherbfield. The sites studied for Tertiary palynology are shown in Fig. 1. Southeastern Australia has been studied most intensively but most ofthese studies present a disjointed sequence which must be pieced together for a coherent history. Dating is usually achieved by correlation with adated sequence as most sediments cannotbe dated directly and there is little independent evidence ofthe age. There are a few sites on the continental shelfand in the deep sea that yield pollen which must have come from the vegetation on land. These sites present a more continuous sequence that is well dated by independent evidence from marine foraminifera. Figure 1 shows that there are large areas for which there is noevidence. This review presents the climatic evolution ofsouthern Australia, which has arela- tively goodrecord, forcomparison with northern Australia where the evidence is sparse. SOUTHERN AUSTRALIA The climate over most ofsouthern Australia is temperate, becoming subtropical in the north and dry continental, semi-arid and arid in the central and inland regions (Foley 1954). Patches of rainforest are found along the east coast and in Tasmania but Eucalyptus forests and woodlands are dominant over most ofthe area. Open shrublands andgrasslands are found in the semi-arid and aridregions. Most of the evidence comes from southeastern Australia and this region covers Tasmania, Victoria, New South Wales and the southeast of South Australia. There are a few sites in central and southwest Australia. Palaeovegetation In southeastern Australia, the Lachlan River Valley on the western Slopes of the Great Dividing Range (H-F-C on Fig. 1) has an almost continuous record from late Eocene to the Pleistocene (Fig. 2). The vegetation was rainforest with abundant Nothofagus throughout the Oligocene, arguably the wettest period in the Tertiary. In the late Oligocene-early Miocene, Nothofagus declined, but the vegetation was still predominantly rainforest (Martin 1987). The mid Miocene was atime ofprofound change when rainforests were decimated and myrtaceous forests became predominant. The carbonised particle orcharcoal content is greater inthe myrtaceous pollen assemblages than in the rainforest assemblages, show- ing thatperiodic burning had become part ofthe environment, thus inferring that the vegetation was sclerophyll forests and not rainforest, for the latter rarely burns. This indirect Proc.Linn.Soc.n.s.w., 119. 1998 118 TERTIARYCLIMATEINAUSTRALIA MelvilleIs., , George • SitesstudiedforTertiarypalynology LalrobeValley (•) Deepseasites studiedforpalynology \ ~Tasmania Figure 1. Sites studiedforTertiary palynology. The sites usedforreconstructions in Fig. 2 are: H, Hillston; F, Forbes; C, Cowra, all ofthe Lachlan River Valley; A, the Aquarius well on the continental shelf; D, Darling River;N,NorthernNewSouthWales;W,westernregionofsoutheasternAustralia. method ofidentifying eucalypt vegetation is necessary as Eucalyptus pollen is difficult to distinguish from other myrtaceous pollen types (Martin 1987). Some rainforest taxa are present in these myrtaceous assemblages, hence the interpretation of wet sclerophyll forests. There is a hiatus of non deposition and/or erosion in the late Miocene, thought to correspond with the late Miocene low sea level (Fig. 2). In the late Miocene-early Pliocene, there is a briefresurgence ofrainforest, followed by a return to wet sclerophyll forest. In the late Pliocene, the tree taxa declined and the vegetation became more open, viz, woodlands, grasslands and herbfields (Martin 1987). When the records from other sites in southeastern Australia are compared with that of the Lachlan River Valley, the trends are generally similar but with some local differ- ences. In Tasmania, the palynological record is limited (Kershaw et al. 1994), but there is a rich macrofossil record (Hill 1992; Carpenter et al. 1994). The Tertiary vegetation con- tained abundant rainforest taxa which persisted into the Quaternary when taxa common in modern sclerophyllous heathlands and woodlands became increasingly common. The late Quaternary vegetation was mainly Eucalyptus, Asteraceae and Poaceae, but Nothofagus and other rainforest taxa were still present as well. Thus rainforest has had a Proc.Linn. Soc.n.s.w., 119. 199X H.A.MARTIN 119 CO _l >LU 111 _l < LU CO /I c<± % ¥ O <LU LU LX CO C0=") LU CD O 5 m C®«CoODS-rySni—C°CQ<DOn. cot: £8.8 CoL CO £ — o.E LU C O CO CD fSg L>U BNaoond 3N30OIW 3N3009I10 3N30O3 sjBeAuoi||!i/\| Figure2. Reconstructionsofthe vegetationandclimate from lateEoceneinto Pleistoceneforthe Lachlan River Valley,fromHillstontoCowra. SeeFig. 1 forlocation. Presentdaylevelsofprecipitationarefor: H,Hillston;F, Forbes; C, Cowra. Oxygen isotope temperatures are from Shackleton and Kennett (1975) and global sea levels fromHaqetal.(1987).Note:Thesurfaceseatemperaturesarenotcorrectedforcontinentalicevolume(seetext). Proc.Linn. Soc.n.s.w., 119. 1998 120 TERTIARYCLIMATEINAUSTRALIA continuous presence through the Neogene and Quaternary in Tasmania (Carpenter et al. 1994; Kershaw etal. 1994). The early-mid Miocene ofthe Latrobe Valley, southeast Victoria (Fig. 1) has abundant Nothofagus and other rainforest taxa. Casuarinaceae and sclerophyllous taxa are present as well. The rainforesttaxa are also well represented in the late Miocene but have disappeared in the late Pliocene-early Pleistocene, when Eucalyptus, Asteraceae and Poaceae become abundant (Kershaw et al. 1994; Blackburn and Sluiter 1994). In the Southern Highlands (Lake George and other sites in the region, Fig. 1), rainforest is well represented in the early-mid Miocene and extends into the late Pliocene, after which Asteraceae and Poaceae become prominent (Kershaw et al. 1994). In northern New South Wales (N on Fig. 1), rainforest is well represented in the mid-late Miocene. In the Pliocene, there is very little Nothofagus or 'wet' rainforest, but the gym- nosperm taxa, and especially Araucariaceae ('dry' rainforest) are well represented. Asteraceae and Poaceae become abundant in the Pliocene-Pleistocene (Kershaw et al. 1994). The western region of southeast Australia (W on Fig. 1) shows similar patterns, with rainforest well represented in the late Oligocene to early-mid Miocene and dimin- ishing in the Pliocene (Kershaw et al. 1994). Late Oligocene-early Miocene assemblages along the Darling River, the most northwesterly part ofsoutheastern Australia (D on Fig. 1), have a diverse rainforest flora, but in low frequencies and the vegetation was mainly sclerophyll, with some pockets of rainforest, probably along the river valleys, the habitats with the most favourable mois- ture relationships. Asteraceae and Poaceae become common in the Pliocene-Pleistocene (Martin 1997). The sites studiedfrom southwestAustraliaand southerncoastal South Australia (Fig. 1) are mainly Eocene in age, with one Pliocene site. The Eocene palynofloras have abun- dantNothofagus and other rainforest taxa (Macphail et al. 1994) and are generally similar to those in southeastern Australia. By the Pliocene, the vegetation had become sclerophyllous forests orwoodlands dominatedby Casuarinaceae orMyrtaceae (Bint 1981). Palaeoclimate The precipitation deduced from the climatic requirements of comparable present- day vegetation (see Martin 1987) is shown in Fig. 2. The early Oligocene was a time of very high precipitation, over 1800 mm, with high and constant humidities. Precipitation declined somewhat in the Oligocene, and in the early Miocene, it was probably closer to 1500 mm, the lower limit for widespread rainforest in New South Wales today. In the mid-late Miocene, precipitation decreased to 1000-1500 mm, the limits for wet sclerophyll, and there must have been a marked dry season to allow burning on aregularbasis. mm In the late Miocene-early Pliocene, precipitation increased to over 1500 at the time mm of the rainforest revival. Subsequently, it decreased to below 1500 and in the late Pliocene-Pleistocene, to about 500-800 mm (Fig. 2). The present day levels ofprecipita- tion are shown on Fig. 2. for comparison, and they are less than those for the late Pliocene-early Pleistocene. When the palaeovegetation of the other sites in southeastern Australia are compared with that ofthe Lachlan River Valley, they suggest that it was wetter in Tasmania and the southern highlands and drier in the northwest of the region, just as it is today (Martin 1986; 1990a). Temperatures may be deduced from the palaeovegetation in a similarway to that of precipitation, but these two factors are not independent. When temperatures are higher, there is more evaporation and the climate is more humid. Once the climate becomes colder, it is also drier. The oxygen isotope record from deep sea cores (Shackleton and Kennett 1975) has been used to estimate surface sea temperature but it also indicates continental ice volume. In the Neogene and Quaternary, however, the record is influ- Proc . Linn. Soc. n.s.w., 119. 1998 H.A.MARTIN 121 enced by both factors and corrections for ice volume are applied (Feary et al. 1991; Isern et al. 1996). Figure 2 presents the surface sea temperatures from Shackleton and Kennett (1975). Temperatures are high in the late Eocene, probably the highest for the Tertiary, decreasing in the Oligocene, when the Antarctic ice cap was developing, and increasing in the early-mid Miocene. Temperatures show a marked decline in the late Miocene, with a subsequent rise in the late Miocene-early Pliocene, followed by a sharp decline in the late Pliocene. Temperatures of the Pleistocene fluctuated with the glacial/interglacial cycles. The decline in temperatures through the Tertiary parallels the build up of ice on Antarctica (Kennett 1993). Precipitation and temperatures (Fig. 2) are roughly parallel, especially from the Miocene on. Warmer times were also wetter and conversely, colder times are drier, as expected. In the Australian context, precipitation has been the overriding control, hence a clear, unambiguous temperature signal is not evident in the record from the palaeovegetation. The precipitation record in Fig. 2 suggests that there are long periods of relative climatic stability and short periods ofdramatic change. These periods ofchange coincide with some worldwide events which have a profound impact on climate. Global changes in sea levels (Haq et al. 1987) are also shown on Fig. 2. When sea levels are high, the continental shelves and low lying areas on land are flooded. These shallow seas warm up and evaporation is high, hence precipitation is increased. When sea levels are low, the continental shelves are exposed, the land is well-drained and the deep seas bordering the shelves arecold, henceevaporation and precipitation areboth low. The record ofsea level, temperatures and precipitation show roughly similartrends from the Miocene on. There are high levels in the mid Miocene, low levels in the late Miocene, with another high in the early Pliocene. The peaks do not coincide exactly, for there may be an apparent lag due to some specific local factor or imprecise dating. The parallel trends do not hold as well in the latest Tertiary because the ice volume on Antarctica had become a significant factor. The temperature curve from Shackleton and Kennett (1975) was selected because it is the record closest to Australia (see Fig. 2) and hence has most relevance to southeastern Australia, butthis curve is essentially similarto that ofSavin et al. (1975) for the Atlantic and Indian Oceans, and reflects general global trends in climate. In these global records of sea levels and oxygen isotope surface sea temperatures, the periods that standout are as follows: 1. High sea levels and warmer surface sea temperatures in the mid Miocene and the late Miocene-early Pliocene. 2. Low sea levels and cooler temperatures in the mid-late Miocene and again in the late Pliocene-early Pleistocene. The Quaternary glacial/interglacial cycles commenced in this latter period. These two periods mark substantial decreases in both temperature andprecipitation. The times ofmarkedchange in the palaeobotanical record coincide with these global changes in sealevel and temperatures. The periods oflow sealevels anddecreasedtem- peratures, with accompanying lowerprecipitation, were devastatingto the rainforest. NORTHERN AUSTRALIA The northern part ofAustralia is tropical and the coastal strip subequatorial (Foley 1954) and experiences summer monsoonal rains (Leeper 1970). Rainforest is found in patches along the north and east coast and is best developed in northeast Queensland, but overall, rainforest occupies only a relatively small area. Most of the region is covered with open forest, woodland and/or shrubland (Specht 1970). Proc.Linn.Soc.n.s.w., 119. 1998 122 TERTIARYCLIMATEINAUSTRALIA Northeast Queensland Deep sea sites off Cairns and Mackay respectively (Fig. 1) have good records of late Miocene-Pleistocene palynofloras. The vegetation on the land nearest both sites was predominantly araucarian rainforest and casuarinaceous forest (Martin and McMinn 1993). The casuarinaceous pollen could indicate either Gymnostoma, a rainforest taxon, or CasuarinalAllocasuarina, sclerophyllous taxa (Kershaw 1970). There is, however, a suite ofother sclerophyllous taxa, hence there must have been some sclerophyll vegetation. There is some cyclical variation between the araucarian and casuarinaceous vegetation, but no marked change. The pattern of change seen in southeastern Australia is not evident here, and extensive rainforest continues into the Pleistocene, until 120,000 years ago when the araucarian rainforest declines (Kershaw et al. 1993). Today, the Cairns district is the wettest region of Australia, and it may be that in the Neogene, the rainfall was well above the lower limits for rainforest, such that fluctua- tions did not provoke the decline in rainforest seen elsewhere in Australia. These deep sea sites collect pollen principally from the coastal vegetation which would have been closest to the site, whether the sea level was high or low. In contrast, a site on land would register the migration of the vegetation zones as they moved in unison with sea levels (Martin and McMinn 1993). A site on the Atherton Tablelands, inland from Cairns, is thought to be Pliocene- Pleistocene in age (Kershaw and Sluiter 1982). Podocarpus, Nothofagus, Casuarinaceae and Myrtaceae were at times prominent in the vegetation. Eucalyptus is not separated from rainforest Myrtaceae, and there is a wealth of low frequency rainforest taxa. Araucarians were minimal (Kershaw and Sluiter 1982). This assemblage from the tablelands reflects different vegetation to that on the coast, but it is essentially rainforest. The A.O.G. Aquarius No. 1 well on the continental shelf (A on Fig. 1), has a good Neogene sequence (Hekel 1972). Araucarians, Casuarinaceae and Myrtaceae were common in the vegetation, the latter in contrast to the deep sea sites where Myrtaceae is minimal. Rainforest was aconsiderable part ofthe vegetation here also. Rainforest, and hence a relatively high precipitation, was maintained throughout the Neogene in northeast Australia when rainfall was decreasing over most ofthe conti- nent. Environments favourable for rainforest have continued to the present day and the Cairns region is still the wettest part ofAustralia today (Leeper 1970) and has the most extensive tracts ofrainforest. Feary et al. (1991) have reconstructedthe changes in palaeotemperatures using surface sea temperature derived from the Tertiary oxygen isotope record corrected for ice volume and the decreasing latitude with continental drift (Fig. 3). For the Great Barrier Reef, temperatures were subtropical in the Eocene, temperate in the Oligocene, subtropical in the Miocene, and finally becoming tropical in the Pliocene-Pleistocene when the Great Barrier Reefentered tropical latitudes (see Fig. 3) (Feary et al. 1991). On average, temperatures were above the minimum required for tropical reefgrowth (20°C) from the middle Miocene to Holocene, except for intervals in the late Miocene-early Pliocene, when the temperatures fell to between 18° and 20°C repeatedly (Isern et al. 1996) Northwest Australia There are three Tertiary palynological sites in northwest Australia (Fig. 1). The two on land are ?Eocene in age (Truswell and Harris 1982) and contain some rainforest taxa found in deposits ofa similar age in southeastern Australia. There is also an almost continuous sequence from late Miocene to the Recent in a deep sea site, off Port Hedland. There, the pollen would have originated from the land to the south, from Port Hedland and southwards, as there is a north-south canyon down the edge ofthe shelfwhich acted as a funnel (Martin and McMinn 1994). In the late Miocene the vegetation was casuarinaceous forests and there were no Proc.Linn.Soc.n.s.w., 119. 1998 < H.A.MARTIN 123 25-. o o CD 13 *-• CO %_ CD Q. E CD *-• CO CD CO COD CO 13 CO T3 CD 4— CO E '*-> CO LU Time (million years) Figure 3. Inferred surface sea temperature for northeast Australia (from Feary et al. 1996). The upper curve representsthe northernendofthe GreatBarrierReefandthe lowercurve, the southernend. Forfurtherexpla- nation,seetext. unequivocal rainforest taxa. The spores of ferns, bryophytes etc., indicative of damp habitats, are minimal, hence the climate was relatively dry. The casuarinaceous forests decline and are replaced by grasslands. Acacia is present with relatively high frequencies (for Acacia) which is always under-represented, suggesting that it was an impor- tant taxon in the vegetation. The chenopod type, pollen of Chenopodiaceae, is indicative ofarid vegetation and increases in the Pliocene and Pleistocene. There is the minimal Myrtaceae pollen, hence minimal Eucalyptus and the lack of eucalypt dominated vegetation. The vegetation of the region today has large tracts where Acacia shrublands and hummock-tussock grasslands are dominant. This deep sea site probably reflects the development of the hummock-tussock grasslands (Martin and McMinn 1994; Specht 1970). A remarkable macrofossil assemblage from Melville Island (Fig. 1) is, unfortunate- ly, undatable. Cupressaceae, Grevillea, several other taxa of Proteaceae and Melaleuca indicate a non-rainforest community which probably had a seasonal climate (Pole and Bowman 1996). When compared with southeastern Australia, the lack of rainforest and earlier increase in grasses suggests that it was drier in the northwest. Today, the northwest is much drier than the southeast, hence in general terms, climatic gradients in the Tertiary were parallel to those oftoday. Proc.Linn.Soc.n.s.w., 119. 1998 . 124 TERTIARYCLIMATEINAUSTRALIA CENTRAL AUSTRALIA Most of the sites in Central Australia, around Lake Eyre and in the Northern Territory are late Paleocene-late Eocene in age. The vegetation was mainly forests and herbaceous swamps. The rainforests were richly diverse in gymnosperms and there was some Nothofagus, but the latter was not as abundant as in coastal regions. There was a rich angiosperm flora also, and proteaceous taxa were prominent (Wopfner et al. 1974; Truswell and Harris 1982; Alley 1985; Sluiter 1991). Fossil leaf floras with small-sized leaves and an absence of drip-tips indicate some sclerophyllous vegetation (Greenwood et al. 1990, Christophel et al. 1992). Silicified moulds and casts of the fruits of Eucalyptus; Angophora, Leptospermum, Melaleuca/Callistemon and Calothamnus (Lange 1978) add to the sclerophyllous element. There was a mosaic with rainforest in the wetter habitats on the floodplains and along the watercourses with the sclerophyll vegetation on the drierand more nutrientdeficient interfluves. Palaeobotanical material younger than the late Eocene is rare in Central Australia. Once the climate became drier, the swamps etc, required forpreservation, became scarce. When pollen is deposited in a swamp, it must be buried deep enough to escape the destructive effects of a fluctuating water table. The deep weathering, so common in inland Australia, has undoubtedly destroyed much ofthe palaeobotanical evidence. Alate Oligocene-early Miocene assemblage fromLake Frome contains rainforestgym- nosperms, a little Nothofagus and abundant swamp taxa. Grass pollen is rare (Martin 1990b). (Extensive grasslands have been reported from this assemblage, the result of an incorrect identification. This topic isdiscussed furtherbelow.) Atthis time, LakeFrome was afreshwa- terlakewithawide, swampyborder(Callan 1977), withrainforestinthehinterland. A ?Miocene assemblage from the Ti Tree Basin, northwest of Alice Springs has abundant Nothofagus, swamp taxa and some grass pollen (discussed further, later) (Kemp 1978; Truswell and Harris 1982). An ?early-mid Miocene assemblage at Lake Hydra, near Lake Eyre, has minimal rainforest gymnosperms (araucarians and podocarps), the sclerophyllous Casuarinaceae and Eucalyptus, and abundant swamp taxa (Martin unpubl.). Thus there was more sclerophyllous vegetation and limited rainforest in Central Australia in the ?early-mid Miocene, when compared with southeastern Australia, and grasses wererare. Evidence about the vegetation ofthe late Tertiary is very limited, but afew records exist. An ?early Pliocene assemblage at Lake Frome has abundant Casuarinaceae, rare Eucalyptus and a diversity of swamp taxa. There are no rainforest taxa present (Martin 1990b). The vegetation was thus entirely sclerophyllous. A ?late Pliocene-Pleistocene assemblage in the Lake Eyre Basin has low frequencies ofCasuarinaceae and Myrtaceae, the only possible trees, and abundant Asteraceae (daisies), the chenopod type (saltbush- bluebush), Poaceae (grasses) and Cyperaceae (reeds). The vegetation was open shrublands (Martin unpubl.), not unlike the present day arid zone vegetation (Sluiter and Kershaw 1982). The record in Central Australia shows the following progression when compared with southeastern Australia: 1 Late Paleocene-late Eocene: Variedrainforests were predominant andthere were patches ofsclerophyll vegetation. The climate was somewhat drierthan that in the southeast. 2. Late Oligocene-mid Miocene: Sclerophyllous vegetation was well represented but rainforest was still present in the landscape, though much less than in southeast Australia where it was predominant. Swamp taxa were common. 3. ?Early Pliocene: Sclerophyllous vegetation predominated with little or no evidence of rainforest. Rainforest persisted in the more favourable habitats in the southeast. 4. ?Late Pliocene-Pleistocene: The vegetation had become open shrubland, with a significant herbaceous content and more like the arid vegetation of today. Forests were declining in the late Pliocene ofthe southeast. I.I\N. SOC. N.S.W., 119. 1998