1301.0 - Year Book Australia, 2006  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 20/01/2006   
   Page tools: Print Print Page


Dr Mike Smith, Director of Research and Development, National Museum of Australia, Canberra. (Endnote 1)

‘Deserts’, said Isaiah Bowman (Director, American Geographical Society), ‘are no more alike than mountains or plains' (Bowman). In art, film and literature Australian deserts have often been depicted as timeless lands, as ‘a featureless tract of eternity in which nothing had changed or would change’ (Haynes). However, over the last 30 years, research into the archaeology and Quaternary environmental history of Australian deserts has revealed not only their distinctiveness but also their rich history (Veth et al.). A comparative perspective shows the Australian deserts are unlike any other in combining a unique evolutionary history with a comparatively high biomass, high climatic variability and an arid region on a continental-scale. They also have an extraordinary human history. Central Australia has been occupied for 35,000 years, at least as long in fact as modern humans have occupied western Europe (images S7 and S8).

S7: Rock art in the Levi Range, central Australia. Photograph by Mike Smith.
S7 Rock art in the Levi Range, central Australia. Photograph by Mike Smith.

S8 Millstone for grinding grass and acacia seeds. Seed foods were an important staple in the Australian deserts. Photograph by Mike Smith.
S8 Millstone for grinding grass and acacia seeds. Seed foods were an important staple in the Australian deserts. Photograph by Mike Smith.


Deserts are one of the world’s major habitats, forming large bands of drylands along the tropics in both the northern and southern hemispheres, and covering approximately 20% of global land area (Middleton et al.). Collectively the major desert regions in the southern hemisphere make up less than a quarter of this area - only about 500 million square kilometres (sq km) - but they are much more diverse in character than those in the northern hemisphere. They range from the hyperarid Namib (Southern Africa) and Atacama (Chile) deserts, to the arid grasslands or savannas of the Kalahari, the continental dunefields of Australia, and the high-altitude dry Puna (Argentina).


The first thing that strikes you about the Australian deserts is their continental scale reaching from east to west, more than 2,000 km across (map S1). Collectively, they form the largest arid region in the southern hemisphere, covering 3.5 million sq km of desert uplands, salt lakes, stony desert, sand plain, and dune fields (image S9). If the semi-arid zone is included, the area of Australia's desert region exceeds 5 million sq km, or 70% of the total Australian land mass (Taylor). Like the Kalahari, the Australian deserts are well vegetated and only moderately arid. The arid core of the continent centres on Lake Eyre (in the far north of South Australia). It is a vast salt lake of almost 9,500 sq km, where annual rainfall averages about 100 mm. Out from this core, average annual rainfall increases from 250 mm in the south to 500 mm in the north. However, one of the distinctive features of Australian deserts is extreme variability in rainfall (20 -40% more variable than regions of comparable rainfall elsewhere). For instance, Alice Springs (Northern Territory), with a median annual rainfall of 259 mm, received only 54 mm of rain in 1985, but a staggering 903 mm in 1974. (More information is provided in 'Variability in Australian desert rainfall' in Climatic aspects of Australia's deserts.)

Two geographical features merit special mention. The first is that the heart of the Australian continent is ringed by extensive dune fields, where the sand ridges align with the dominant wind patterns of 20,000 years ago (images S10 and S11). In 1845, explorer Charles Sturt was the first European to encounter these: ‘Ascending one of the sand ridges I saw a numberless succession … rising above each other to the east and west … A kind of dread came over me ... It looked like the entrance into Hell’ (Sturt). More than 1.6 million sq km of the Australian arid zone is covered by aeolian sand, either sand plain or dune fields. This sandy mantle is from one to ten metres thick, compared to the great depth of sand (100-200 metres) in much of the Kalahari or eastern Sahara. One of the implications for people is that Australian deserts have a network of small wells, rock holes and soakages. Although these yield only small amounts of water, it is often enough to provide people with a base for exploiting the resources of the surrounding desert.

The second feature to mention is Lake Eyre. This playa lake, which is the terminus of a vast internal drainage system, dominates the eastern half of the arid zone, and has been the focus of the largest sustained program of research into the Quaternary environmental history of the desert. The seasonal rivers, ephemeral lakes, pans and neighbouring sand dunes that make up the Lake Eyre Basin, form 200,000 sq km of riverine desert, and give the
landscapes in this part of the arid zone quite a different character to the sand dune and range country in the Western Desert (Endnote 2) and central Australia (Endnote 3) These river systems can carry large volumes of floodwater into the desert from elsewhere.

S9 Like most desert lakes, Lake Amadeus in central Australia is a vast salt pan. Photograph by Mike Smith.
S9 Like most desert lakes, Lake Amadeus in central Australia is a vast salt pan. Photograph by Mike Smith.

S10 The heart of the Australian continent is ringed by vast dune fields, such as the Simpson Desert. These sand ridges are aligned with the dominant wind patterns of 20,000 years ago. Photograph by Mike Gillam.
S10 The heart of the Australian continent is ringed by vast dune fields, such as the Simpson Desert. These sand ridges are aligned with the dominant wind patterns of 20,000 years ago. Photograph by Mike Gillam.

S11 Spinifex and dunes, near Lake Amadeus, central Australia. Photograph by Mike Gillam.
S11 Spinifex and dunes, near Lake Amadeus, central Australia. Photograph by Mike Gillam.


Although all of the southern hemisphere deserts are long-standing features of the environment, and took shape during the Miocene (24 million to 5 million years ago), they have responded to global and regional climate change during the Quaternary (the last 1.8 million years). The late Quaternary histories of the southern deserts show that all have seen periods of enhanced rainfall, fluvial activity, groundwater discharges and greater biological activity in the past (Smith et al.). In broad terms, these deserts show three sorts of changes. These are:
  • Changes in intensity of aridity
    Within the deserts, the degree of aridity has waxed and waned in response to shifts in global climate and weather systems. Generally, this has involved variations about an arid mean rather than transformation of the desert environment into sub-humid savanna. For instance, the climate of central Australia today probably represents the most favourable period in this desert since the end of the last interglacial (about 120,000 years ago) whereas the peak aridity of the last glacial maximum (about 20,000 years ago) marks the harshest conditions yet identified, with parts of the interior subjected to hyperarid conditions for a protracted period.
  • Changes in extent of the deserts
    The boundaries of these drylands have not been stable. The southern deserts have increased in size or contracted depending on shifts in global climate. In fact, the clearest evidence for environmental change is often from the margins of these deserts. For instance, fossil linear dunes occur well beyond the margins of the Australian arid zone, extending in southern mainland Australia into north-eastern Tasmania, and beneath King Sound in the north of Western Australia. In Australia, the last glacial maximum was not just a time of more intense aridity, but also a period when deserts were more extensive.
  • Changes in favourable patches within deserts
    Even within an arid or hyperarid landscape, external factors may selectively affect parts of the landscape. For instance, the reactivation of dry lakes in the Willandra region (south west New South Wales) or of the rivers feeding Lake Eyre (e.g. Coopers Creek), reflects changes in winter or summer rainfall systems outside the arid zone, and the transport of floodwaters into these regions. For instance, Lake Mungo (one of the lakes in the Willandra system) held water at various times between 15,000 -55,000 BP (Endnote 4), at a time when the surrounding landscape was more arid than today, because colder drier conditions produced more effective runoff in the southern highlands.

Around 313 million people (or about 13% of the world’s population) currently live in the world’s arid zones - about four million in southern hemisphere deserts (Middleton et al.). In 2001 about 180,000 people lived in the centrally located arid zone of Australia. The Australian semi-arid zone supported a further 394,000 people. Overall the desert region of Australia is occupied by less than 600,000 people (or fewer than 3% of the total population) (Taylor).

All of the southern hemisphere deserts have remarkably long records of human settlement: more than 60,000 years in southern Africa, at least 35,000 years in Australia and about 13,000 years in South America. Current archaeological evidence suggests that the southern deserts were explored and colonised as part of the dispersal of modern humans across the globe. In southern Africa, the first sustained use of the Kalahari took place about 100,000 years ago at about the same time that evidence for anatomically-modern and behaviourally-modern humans appears in the archaeological record. In Australia and South America, initial human movement into the deserts took place as part of dispersal across new continents, by people who had already demonstrated the capacity to undertake successful sea crossings (Australia) or cross extreme high latitude cold environments (North America).


Deserts are difficult environments for people because of their low biomass and scarcity of critical resources such as water, fuel, plant foods and game (though these factors vary considerably across the southern deserts). They are also environments where resources are patchy and highly variable in both time and space:
  • In Australian and southern African deserts, where there are significant plant and animal resources thinly distributed throughout the desert, it is the distribution of watering points that determines which parts of the desert landscape people can reach and where and when they can harvest available resources, whether these be stone, ochre, plant foods or game.
  • Both the spatial distribution of these waters and the productivity of the surrounding country vary with rainfall. Rainfall events create unpredictable pulses of biological productivity, separated by long dormant periods. In deserts, much of the ecosystem is geared towards this pattern of ‘boom and bust’ or ‘pulse and reserve’ (Noy-Meir).

All of the southern hemisphere deserts were successfully settled by hunter-gatherers. For historic groups, the key adaptations for living in deserts were behavioural and social, rather than technological. They included high residential mobility, broad-spectrum foraging, a high degree of organisational and technological flexibility and intimate knowledge of the dynamics of the landscape. One other factor is also important here. Human dispersal is likely to be constrained by the difficulty of maintaining viable social networks when population levels are very low (as in most deserts). Historically, the Australian desert had some of the lowest population densities on record for human populations (as low as one person per 100-200 sq km). Under such conditions, effective social networks are important.


The Namib and Kalahari preserve the earliest traces of people living in the southern deserts. Early Stone Age Acheulian hunters followed game and water into these areas about 300,000 -400,000 years ago, wherever pans, springs or floodwaters provided accessible routes into the deserts. By 60,000 -90,000 years ago (during the Middle Stone Age), there is good evidence for establishment of a resident hunter-gatherer population in the southern Namib and in the northern part of the Kalahari Desert (the ‘Middle Kalahari’ basin). The first human movements into the Australian deserts took place sometime before 35,000 years ago, probably as part of initial peopling of the continent. By 45,000 years ago people were present on both the northern and south-eastern margins of the Australian arid zone. By 30,000 BP, small groups of highly mobile hunter-gatherers were using pockets of country across the interior of the continent, from central Australia to the Pilbara (in the north of Western Australia), and from Lake Mungo to the southern Kimberley (Western Australia). Settlement of American deserts also took place in the context of initial human colonisation of a continent. Hunting groups moved into the Andes mountains as soon as the glaciers retreated. They followed the puna steppe grasslands south into the Atacama region, and had moved down into the deserts on either side of the Andes by 10,500 -11,000 BP (13,000 years ago) - about the same time that Clovis hunters were moving into North American deserts such as the Mojave.

This picture is complicated by the dynamic environmental history of these regions - were they deserts when people first arrived? - as well as uncertainty about the nature of the earliest occupation in these extreme environments - are we looking at successful colonisation of these deserts, or simply intermittent visits by people? Successful settlement of a new region (‘colonisation’) may well have been preceded by an exploratory or pioneer phase (‘dispersal’).


What sort of environment did people find when they entered the Australian deserts? Between about 100,000 and 13,000 years ago, the interior of the Australian land mass was more arid than present. The exception is the south-eastern section of the arid zone, where rivers and lakes in the Darling Basin and Willandra region (New South Wales) were more active during between 55,000 and 15,000 years ago. Elsewhere in the Australian desert region, most dates for enhanced fluvial activity (such as along channels feeding Lake Eyre) or high lake levels (e.g. Lake Eyre, Lake Gregory (far north Western Australia), and Lake Woods (Northern Territory)) centre on 100,000 years ago, or earlier in the last interglacial. At this time, there were significant water bodies in playa lakes in the northern and eastern parts of the arid zone, but no evidence for palaeolakes in the western half of the desert. The last deep-water phase of Lake Eyre in central Australia ended 60,000 years ago, several millennia before humans arrived on the scene. The period between 60,000 and 24,000 years ago saw some reactivation of rivers and lakes outside the desert, and on its margins, but seems to have had only limited impact in central Australia, except perhaps at Lake Frome (South Australia).

There are several implications for the human ecology of the arid zone:
  • the large palaeolakes, and major fluvial activity along arid zone rivers, both ended well before the first human movements into the desert (except in the Darling and Willandra regions, and at Lake Frome)
  • apart from the Darling and Willandra lakes, the palaeolakes are unlikely to have ever held significant resources for people. Lake Frome, Lake Eyre, Lake Amadeus and Lake Lewis (Northern Territory) were saline or brackish bodies of water, notably without significant accumulations of fish bone or freshwater mussel shell along their shores
  • current (limited) vegetation data are consistent with an arid vegetation over the last 100,000 years, indicating that terrestrial dryland resources would have been the mainstay of early settlement in the desert
  • potable water may have been more widespread before 30,000 years ago, giving more flexibility to annual and seasonal subsistence movements in the desert

By 30,000 years ago, people were using a range of habitats across this huge region including:
  • montane and piedmont areas, such as the central Australian ranges (Puritjarra 32,000 BP; Kulpi Mara 29,500 BP) and the Pilbara (Newman rock-shelter >26,000 BP)
  • sandy deserts (Serpents Glen >23,600 BP)
  • karst landscapes, such as the Nullarbor Plain (Allens Cave 39,800 years ago)
  • arid littoral regions such as the west coast of the continent at North West Cape (Mandu Mandu 34,000 BP; Pilgonaman Creek 32,000 BP)
  • riverine or lake habitats in the south-eastern sector of the arid zone (Lake Mungo, 45,000-50,000 years ago)

There are also several early sites on the margins of the desert, in the southern Kimberley (Carpenter’s Gap and Riwi, both dated 45,000 years ago), in an area incorporated into an expanded arid zone during the drier climates of the last glacial.

Most of these sites show repeated use over several millennia. Where there are data on economic orientation, they indicate these early groups were generalised foragers exploiting reptiles, small macropods, and emu eggs (supplemented with marine or lacustrine resources where available). In the Willandra, local exploitation of fish and shellfish probably reflects small-scale seasonal use of lacustrine resources, supplemented with terrestrial foods. At Puritjarra, in the west of central Australia (one of the few desert sites to have been investigated in detail) the evidence indicates sustained occupation of the central desert from about 35,000 years ago, with exploitation of regional ochre mines and local stone sources beginning around this time.

There are few indications, in these data, that archaeologists have yet uncovered the ‘pioneer’ phase associated with initial movements into the Australian deserts. Given that human dispersal across the continent took place before 45,000 years ago and that the northern margins of the arid zone were settled soon after, we could expect that the first movements into these deserts took place 35,000-45,000 years ago.


The histories of the southern hemisphere deserts indicate that colonisation was not a discrete event but rather an ongoing process - involving expansion of settlement from core to peripheral areas within deserts, more intensive use of desert resources (especially plant foods), and the gradual elaboration of social and economic landscapes. Human settlement also remained vulnerable to changes in environment, especially changes driven by major shifts in climatic parameters.

Intense aridity during the peak of the last glacial, centred on 20,000 BP, may have created difficult conditions for people in many parts of the arid zone. Palaeoenvironmental recon-structions for the period from about 30,000-19,000 years ago suggest increased seasonal and diurnal contrasts, saline groundwater, strong winds, an intensification of aridity and significantly lower mean annual air temperatures. This is likely to have led to a contraction of settlement in the arid zone, and abandonment of some regions, but researchers differ on the extent, scale and duration of any impact. Parts of the Western Desert and Lake Eyre basin may have been abandoned at this time - or in the period of rising global temperatures immediately after the glacial maximum. The clearest evidence for site abandonment is from Serpents Glen, in the Carnarvon Range (Western Australia), where there is a sterile layer, dating between 24,000 BP and 5,000 BP. In other regions, settlement appears to have continued. The Pilbara and central Australia both saw a reorganisation of land use at this time, with more focussed use of sites near water or in desert uplands. In western central Australia, people appear to have continued to visit Puritjarra rock shelter intermittently throughout the glacial maximum, suggesting that human occupation of the central ranges and the sandy desert immediately west of the ranges continued throughout this period.


The population of the Australian deserts increased again and re-occupation of abandoned regions took place as arid conditions ameliorated from about 13,000 BP. Nevertheless, all of the southern hemisphere deserts contain areas that were not effectively colonised until the last few millennia (such as the nitrate pampa in the Atacama, the gravel plains and mountains of the central Namib, the sand plain of the central Kalahari and some parts of the sandy deserts or continental dune fields in Australia).

In the Australian deserts, the last thousand years appears to have been a period of major change in the Western Desert and in central Australia. There are indications of more sedentary occupation or increased levels of site use at about 1,000 BP, possibly as part of a demographic transition towards higher regional populations in the Australian deserts. Recent work is beginning to refine this picture, showing that changes in settlement pattern may have began as early as 3,000 BP. Even within the last millennium there is evidence for cultural change at a range of geographic and time scales, particularly within the last 500-200 years. Archaeological evidence also shows that many aspects of the technology, economy and subsistence behaviour of historic desert groups are relatively recent, rather than representing features of early desert occupation. Australian desert societies have long histories of development, adaptation and response to life in these extreme environments.


1. This article is adapted from material previously published in Smith, MA & Hesse, P (eds.) 23S: Archaeology & environmental history of the Southern Deserts, Canberra, National Museum of Australia Press, 2005. < Back

2. Includes the Great Sandy, Little Sandy, Gibson and Great Victoria deserts. < Back

3. Includes the Tanami, Simpson and Strzelecki deserts. < Back

4. Before 1950, being the accepted radiocarbon dating reference year. < Back


Bowman, I ‘Desert Trails of the Atacama’, New York: American Geographical Society, Special Publication 5, 1924, p. 60.

Haynes, RD ‘Seeking the Centre: The Australian Desert in Literature, Art and Film’, Cambridge University press, 1998, p. 124.

Middleton, RD & Thomas, DSG (eds), ‘World Atlas of Desertification’, 2nd edition, London, UNEP/Arnold.

Noy-Meir, I ‘Desert ecosystems: Environment and producers’, Annual Review of Ecology and Systematics 4, 1973, pp. 25-51.

Smith, MA & Hesse, P (eds.) ‘23S: Archaeology and environmental history of the Southern Deserts’, Canberra, National Museum of Australia Press, 2005.

Sturt, C ‘Journal of the Central Australian Expedition 1844-1845’, London, Caliban Books, 1984, pp. 73-74.

Taylor, J ‘Population futures in the Australian desert, 2001-2016’, Discussion Paper No. 231, Centre for Aboriginal Economic Policy Research, The Australian National University, Canberra, 2002, p. 4.

Veth, P Smith, M & Hiscock, P (eds.) ‘Desert Peoples: Archaeological perspectives’, Oxford: Blackwell Publishers, 2005.