When humans first set foot in Australia more than 65,000 years ago, they faced the perilous task of navigating a landscape they’d never seen. Now, researchers have used supercomputers to simulate 125 billion possible travel routes and reconstruct the most likely “superhighways” these ancient immigrants used as they spread across the continent. The project offers new insight into how landmarks and water supplies shape human migrations, and provides archaeologists with clues for where to look for undiscovered ancient settlements.
“This is a really compelling illustration of the power of using these [simulation] techniques, at a huge, continental scale, to understand how people navigate landscapes,” says archaeologist Kyle Bocinsky of the University of Montana, Missoula. “It’s impressive, extreme computing.”
Researchers have long puzzled over the routes humans first took to Sahul, the landmass that connected Australia, New Guinea, and Tasmania before the islands were separated by rising seas some 20,000 years ago. Most researchers think ancient people used boats to cross from Asia. But the routes they walked once they reached land have never been obvious.
To reconstruct those early travelers’ potential paths, researchers assembled the first detailed topographic map of Sahul—including past coastal lands now underwater—from satellite, aerial, and undersea mapping data. Next, they turned to a computational technique called least-cost path analysis to calculate optimal walking routes across this re-created landscape. The calculations are analogous to pouring digital water on a virtual landscape and watching where it flows, following the paths of least resistance.
In this case, the researchers created an imaginary traveler—a 25-year-old woman carrying 10 kilograms of supplies—and sent her on billions of journeys across Sahul. Their simulation calculated how many calories she would have burned while walking each route, as well as which paths would have provided a reliable supply of water, based on what they knew about the lakes, rivers, and springs that existed in Australia at the time. The team also simulated what the traveler would have seen as she walked, noting prominent landmarks—such as craggy outcrops or prominent ridges—that she could have used as navigational aids.
It took weeks to run the complex simulations on a supercomputer operated by the U.S. government, says co-author Devin White, an archaeologist at the U.S. Department of Energy’s Sandia National Laboratories who developed several of the computational approaches. “We were grabbing time at night and on weekends,” he recalls. But the number crunching ultimately revealed a network of “optimal superhighways” that had the most attractive combinations of easy walking, water, and landmarks, the researchers reported last week in Nature Human Behavior.
Optimal road map in hand, the researchers faced a fundamental question, says lead author Stefani Crabtree, an archaeologist at Utah State University, Logan, and the Santa Fe Institute: Was there any evidence that real people had once used these computer-identified corridors? To find out, the researchers compared their routes to the locations of the roughly three dozen archaeological sites in Australia known to be at least 35,000 years old. Many sites sat on or near the superhighways. Some corridors also coincided with ancient trade routes known from indigenous oral histories, or aligned with genetic and linguistic studies used to trace early human migrations. “I think all of us were surprised by the goodness of the fit,” says archaeologist Sean Ulm of James Cook University, Cairns.
The map has also highlighted little-studied migration corridors that could yield future archaeological discoveries. For example, some early superhighways sat on coastal lands that are now submerged, giving marine researchers a guide for exploration. Even more intriguing, the authors and others say, are major routes that cut across several arid areas in Australia’s center and in the northeastern state of Queensland. Those paths challenge a “long-standing view that the earliest people avoided the deserts,” Ulm says. The Queensland highway, in particular, presents “an excellent focus point” for future archaeological surveys, says archaeologist Shimona Kealy of the Australian National University.
Bocinsky hopes other researchers will use the technique to map early migration routes across other continents, including North America. But he concedes that “it’s not going to be easy for someone to replicate … if they don’t have a supercomputer.” If they do, they could get more than just an understanding of past human migrations. The muscular digital tool, White says, could help researchers understand how future migrants might flee problems such as war and global warming. “There’s a lot of potential here.”