Sahul

Colonisation of Sahul 70-60 thousand years ago (kya) represents the first great maritime migration undertaken by anatomically modern humans in one of the final phases of the Out of Africa dispersal. Visual connectivity network analyses, agent-based simulations and ocean current modelling reveal that modern humans could follow numerous northern and southern migration pathways into Sahul. Our results support a southern route out of Africa through South Asia with entry into ISEA through the Banda Arc, culminating in an early colonisation of Sahul on the northwest shelf. Our results show multiple colonisation events through other entry points were also probable, and raise interesting possibilities for complex regional migration and population histories.

An understanding of the demographic circumstances and ecological repercussions of the arrival of the first people to Sahul (mainland Australia, Tasmania and New Guinea joined at times of lower sea level) in the Late Pleistocene remains elusive 1-3. Some have previously proposed that the first peopling arose from only small family groups consisting of <150 people 4-7 , while more recent human behavioural ecology models suggest that several hundred people would have been required for long-term survival 7. These estimates are largely speculative, but genomic research and radiocarbon-inferred demographic modelling support the larger values, and imply that population sizes of 1,000-3,000 people were more likely 8-10. Quantifying the plausible demography of the first humans (that is, the rate of population changes relative to regional carrying capacity, and the duration over which populations could have persisted at low density following first arrival) is essential to ascertain the extent to which increasing human populations could have altered their environments. Quantifying demographic transitions can also potentially help to interpret the likelihood of discovering archaeological evidence, given the persistence of small human populations over extended windows of time so long ago 6. The oldest archaeological evidence claimed in Australia dates to 65.0 ± 5.7 thousand years ago (ka) (95% confidence interval (CI)) for the Madjedbebe rockshelter in Arnhem Land 11,12 , and an increasing number of early sites have been reported dating to around or before 47 ka 13-21. We therefore take the broad interval of 65-50 ka as the probable arrival window of people into Sahul. Arrival probably occurred somewhere in the northern regions of Sahul, closest to the islands of Wallacea in the Timor Sea, the Arafura Sea, and across the Torres Strait and Carpentarian Plain during times of lower sea levels 22-27. However, given that only a few archaeological sites from the north have revealed cultural remains within the putative arrival window, and that ancient DNA is poorly preserved in this region of the world, it is difficult to estimate when and where enough people first arrived in Australia to produce one of the longest standing, successful human populations in the world outside Africa-a population that went on to adapt successfully and populate the entire continent over the following several thousand years 10,28,29. Recent studies have modelled plausible routes, and therefore the potential geographic locations that would have supported a successful migration to Australia 22-27 , but the numbers and diversity of humans who first arrived on the continent remain largely unknown 10,28,29. Fundamentally then, the size and migration patterns of founding populations (that is, whether arrival was accidental by a small band of hunter-gatherers, or something larger scale and more complex) directly contributes to our understanding of modern human societies at this time, and how the peopling of Sahul fits into this broader story. To determine the probable range of these unknown demographic conditions, we developed a stochastic, age-structured demographic model for ancient Australians to: (1) estimate the minimum size of a founding population that would be required to avoid a high risk of extinction at the time of first arrival; and (2) calculate the interval and frequency of smaller introductions that would maintain a low probability of extinction over the initial arrival window (65-50 ka). Our model is based on realistic estimates and assumptions of hunter-gatherer demography, as well as a reconstruction of the carrying capacity based on hindcasted estimates of net primary production. We hypothesize that several thousand individuals arriving over a defined period within several centuries were required to The timing, context and nature of the first people to enter Sahul is still poorly understood owing to a fragmented archaeological record. However, quantifying the plausible demographic context of this founding population is essential to determine how and why the initial peopling of Sahul occurred. We developed a stochastic, age-structured model using demographic rates from hunter-gatherer societies, and relative carrying capacity hindcasted with LOVECLIM's net primary productivity for northern Sahul. We projected these populations to determine the resilience and minimum sizes required to avoid extinction. A census founding population of between 1,300 and 1,550 individuals was necessary to maintain a quasi-extinction threshold of ≲0.1. This minimum founding population could have arrived at a single point in time, or through multiple voyages of ≥130 people over ~700-900 years. This result shows that substantial population amalgamation in Sunda and Wallacea in Marine Isotope Stages 3-4 provided the conditions for the successful, large-scale and probably planned peopling of Sahul.