The quest to understand when ancient migration to Australia began

Australia, the continent, wasn’t always Australia as we know it today.

At the end of the Pleistocene – an epoch spanning 2.5 million years and ending just 11,700 years ago – it was part of a far larger landmass fused with New Guinea and Tasmania.

Archaeologists call it Sahul.

And in just a matter of decades, as more and more sites of archaeological interest have been uncovered, scientists have worked to answer Australia’s ancient migration questions.

Until quite recently, many had agreed that Sahul had been occupied by humans for at least 45,000 years, based on genetic studies and archaeological evidence.

How, where and why they arrived has been a source of speculation.

Parts of the ‘how’ are well-supported: it’s clear that water had to be crossed. Ocean separated Sahul from Sunda – the exposed landmass of modern-day South-East Asia.

Yet other questions have persisted: was settlement an accident, or something more deliberately planned? By which route – or routes – was the continent first settled, and when?

As is the nature of western science, further investigation has prompted shifts in thinking, and within the last decade, scientists have found reason to potentially shift the date of Sahul’s first landings even further back in a mind-bogglingly long history of time – potentially by 15,000 years.

And that partly owes to a big dig in the middle of Arnhem Land.

Does Madjedbebe up-date human arrival in Australia?

In 2017, research published in Nature dated evidence of human settlement on Mirarr Country in Arnhem Land at 65,000 years – or conservatively around 59,300 years, given some uncertainty with dating techniques.

Even scientifically, that’s a big update.

The study site was the Madjedbebe rock shelter, which has undergone several excavations over half a century. Even in 1990, a dig at the site uncovered artefacts dated between 50 and 60,000 years.

“And that became quite controversial,” says Professor Chris Clarkson, an archaeologist from the University of Queensland, who co-led the more recent excavations at Madjedbebe.

UQ brokered a deal with the Gundjeihmi Aboriginal Corporation to explore the shelter in 2012. It marked the start of a partnership between researchers and the Mirrar people to timestamp the arrival of humans to the region.

Layers of earth were removed to reach thousands of stone artefacts buried beneath the surface. Among them were stone axes and grinding tools believed to be the oldest known examples of such instruments on the continent, as well as ground ochre that might have once been used in art practices.

The tools themselves can’t be easily dated, but finding them alongside other material allows a kind of ‘dating by association’.

“Organic materials like charcoal from the fireplace or even human or animal bones could potentially be directly dated if there’s still collagen or dateable organic material inside them, which is often the case,” Clarkson says.

“Dating pieces of charcoal, maybe from a campfire next to or surrounding the artefacts, or it might be, in the case of luminescence dating, taking sand grains out of the surrounding sediments.”

Increasingly, scientists are obliged to engage with indigenous peoples where research directly intersects with culture and land. Throughout the process brokered with them, the Traditional Owners of Madjedbebe, the Mirarr, via the Gundjeihmi Aboriginal Corporation, told Cosmos its people had actively participated in the excavations and ensuing study of recovered artefacts.

“Mirarr have been very involved throughout from the stages of drafting and agreeing to the terms of the research through every stage of the fieldwork and ongoing involvement in writing published scientific articles about the research outcomes. Mirarr have an ongoing relationship with several archaeologists to continue investigating Mirarr Country.”

The Madjedbebe study is not without its critics. Some, like Emeritus Professor Martin Williams, a highly-published geomorphologist working across Africa, Australia, India and China have pointed to termite disturbance as an explanation why the Madjedbebe dating was seemingly out-of-sync with other archaeological records. Termites, he and his collaborators have suggested, are capable of disturbing sediment layers such as leading to inaccurate dating of artefacts within sediment layers where luminescence techniques are used.

Others, such as the late desert field archaeologist Dr Mike Smith and colleague Dr Ingrid Ward disagreed with his assessments of the Madjedbebe site.

But where groups of archaeologists back-and-forth over the timestamping of Madjedbebe, other science has supported the findings.

“Without Madjedbebe, it falls over.”

In the last few years, collaborating scientists have sought to plot how early humans might have reached Sahul.

While it had been assumed at a single-entry point to the continent acted as a gateway for people to populate what would eventually become Australia, modelling developed by a team of Australian researchers indicates the existence of potentially two entry points, leading to a series of migratory ‘superhighways’ – plural.

Research into the subject published in recent years has been led by Professor Corey Bradshaw, an ecologist from Flinders University.

He’s an ecologist by training – not an archaeologist – but applying techniques from his work studying biological mechanisms, he set about using complex mathematics to model complex human systems.

These large stochastic models give scientists a tool to predict outcomes while accounting for a myriad of random variations over time.  

“It’s a butterfly effect kind of idea: if you do this, what happens down the track?” he says.

This cross-disciplinary research brought together a mix of geomorphologists, climatologists and archaeologists to review a range of data on ancient migration to Sahul to crunch the probable routes that people arrived on the continent.

“It started off really as what were the more plausible routes through now what we call Indonesia through into Sahul,” Bradshaw says.

“There’s been various modelling attempts, but we hadn’t actually put the human element into that in terms of from a demographic perspective, so trying to put in some of the main concepts of what we call minimum viable population sizes.”

Conceptually, a minimum viable population is what it says on the label: the smallest number of individuals needed for a group of organisms to survive, 95% of the time.

In this ecological scenario, the ecosystem is Sahul. The organism is Homo sapiens.

“In most introductions of any species – think of invasive species – they fail,” Bradshaw says.

“Basically, genetics kicks in, and you get inbreeding, depression, and stochastic events that wipe people out, or individuals out.

“And so the possibility of people just randomly floating in from island to island is quite implausible. So we wanted to put that into context of demography.”

While navigable routes to Sahul existed on paper, Bradshaw and his colleagues sought to test whether they could plausibly sustain a human population.

What they deduced was, rather than a chance colonisation of the continent, Sahul was first inhabited as the result of what Bradshaw calls an extremely organised and directed migration – a view somewhat supported by some critics of the Madjedbebe dating exercise.

“Clearly, it had to evolve quite advanced watercraft,” Bradshaw says.

“They weren’t just falling in the water and being passively taken by ocean currents from island to island. These were directed, well-coordinated, planned and complicated migrations with intention.”

Indigenous arrival huge planned migration

From 120 scenarios modelled by Bradshaw’s group emerged a finding that people were more likely to have arrived on the continent’s north-west region – today called the Kimberley – at least 50,000 years ago.

Another of Bradshaw’s scenarios suggests a wave may have also migrated to Sahul via the western end of what is now New Guinea.

While conjecture among archaeologists exists as to the date of Madjedbebe at around 65,000 years, Bradshaw’s model helps firm the date: his modelled outcome made less sense when the site’s location and date were excluded.

Bradshaw’s model looked at the entire Australian continent, connecting a plausible entry point to dozens of archaeological sites across New Guinea, Australia and Tasmania – effectively playing a game of dot-to-dot at a continental scale.

“If I remove it [Madjedbebe at 65,000 years ago], the entire system falls apart,” Bradshaw says.

Put the site and time back in, and the rate and direction of human movement, which would lead to Australia being fully inhabited from a single entry point within at least five millennia, is far more plausible.

In the genes?

Relatively new to science’s quest to understand migration to Sahul is the study of ancient DNA – trying to quantify, in terms of time, how far back a genetic lineage goes.

“Like with any discipline, it all comes with a lot of caveats,” says Dr Ray Tobler, an evolutionary geneticist from the Australian National University.

That’s because, as with archaeology, the sophistication of dating techniques in genetics has limitations.

Evolutionary geneticists like Tobler lean on mitochondrial DNA (mtDNA) to understand how far back lineages can reach. mtDNA – a circular chromosome found in the powerplant of animal cells – is incredibly well conserved in humans.

“Basically, mtDNA is nice to work with when you’re looking at inferring human or genetic relationships, and then using that to try identify certain aspects of our history, such as when people arrived in Sahul,” Tobler says.

Using mtDNA obtained from hair samples (with consent from living Aboriginal people or their families) geneticists have refined the collective understanding of human movement into Sahul.

“When you’re looking at genetic data, none of it tells you exactly when people arrive at a particular place,” Tobler says.

Instead, researchers have to develop phylogenetic trees – the biological equivalent of a family tree –  which demonstrate genetic relationships; in this case, between mtDNA samples.  

Tracing these trees back to a single common ancestor – a point where groups of humans split genetically – lands somewhere in the range between 42 and 47,000 years ago. Tobler describes this as a “lower boundary” – one that is still an estimate, based on the mutation rate of genes that lead to diversification between human groups.

“If you accept that people were in Australia when these lineages started diversifying, then that’s a lower boundary for the arrival of people,” he says.

“Effectively, what you’re doing is counting the number of mutations that are occurring between these different lineages, and if you’re assuming that they’re occurring at a certain rate, that gives you a time to how long it has to have been to accumulate that amount of diversity.”

To take a helicopter view, two branches of science – archaeology and genetics – agree a group of people have inhabited a single, long-isolated landmass for an unbroken expanse of 45,000 years, at the very least, and probably longer. 

Whichever side you fall on, it’s still an incredible amount of time

Mirarr know their ancestors have used Madjedbebe for such an expanse of time that can’t adequately be described by western concepts of human history.

Many indigenous peoples across the globe will describe their habitation and relationship to an area as being of “time immemorial” – literally for a time not in memory.

For them and other indigenous peoples around the world, the quest for western science to chart humanity’s journey across time and space is, conceptually, quite different to their own.  

But it can add value to sharing their stories.

“The date of 65,000 years, or any dates the archaeological community comes up with, are seen as a Balanda [non-Aboriginal] way of understanding deep time,” representatives for the Gundjeihmi Aboriginal Corporation told me.

“The exact numbers are not hugely relevant, however, there is a clear understanding within Mirarr of the importance of the site from a scientific and international perspective.

“The fact that Balanda methods have ‘proved’ [sic] something Mirarr have always known, increases Mirarr opportunity to share the stories of why their country is important.”

Among the scientists and researchers Cosmos spoke to for this piece, including those not featured, was a view that the time scales being contested around one particular site are themselves at such a distance from modern understanding that it risks undermining the broader point.

That is: the first peoples of Sahul, and later Australia and New Guinea, have occupied these lands for such an extensive period of time as to have become intrinsically intertwined with Country.

“There’s a really long-standing connection between people and place in Australia that is, maybe, unique in the world,” says the geneticist Tobler.

And as archaeologist Clarkson says,  “That is virtually forever.”

During NAIDOC Week (2-9 July 2023), an annual observance in Australia that celebrates and recognises the history, cultures and achievements of Aboriginal and Torres Strait Islander peoples, Cosmos is publishing a series of articles on Australia’s First Peoples and science.