Long-standing questions about the migration of early modern humans in East Asia may finally be answered, thanks to a rare and remarkable journey made in a dugout canoe.

The timing and destinations of the earliest modern human migrations into East Asia are fairly well established. What remains unclear is how these early populations managed to travel between islands separated by dangerous ocean passages.

To address this gap, a research team from Japan and Taiwan, led by Professor Yousuke Kaifu of the University of Tokyo, explored the types of methods ancient people might have used. They also recreated the journey themselves, building canoes with replicas of the tools available at the time.

Archaeological and environmental evidence indicates that roughly 30,000 years ago, humans completed a sea voyage from what is now Taiwan to islands in southern Japan, such as Okinawa—without the aid of maps, metal tools or modern seafaring vessels. To better understand how this journey could have taken place, Kaifu’s team conducted both experimental reconstructions and computer-based simulations.

One of their two recent studies used digital modeling to test how a vessel could cross the powerful Kuroshio Current, one of the world’s most forceful ocean flows. The results demonstrated that a craft built with Paleolithic-era tools and navigational knowledge could indeed manage the crossing. The second study focused on building and trialing an actual canoe, which the team used to paddle more than 100 kilometers between islands, successfully replicating the hypothesized ancient route.

Experimental archaeology at sea

“We initiated this project with simple questions: ‘How did Paleolithic people arrive at such remote islands as Okinawa?’ ‘How difficult was their journey?’ ‘And what tools and strategies did they use?’” said Kaifu. “Archaeological evidence such as remains and artifacts can’t paint a full picture as the nature of the sea is that it washes such things away. So, we turned to the idea of experimental archaeology, in a similar vein to the Kon-Tiki expedition of 1947 by Norwegian explorer Thor Heyerdahl.”

In 2019, the research team built a 7.5-meter dugout canoe named Sugime, carved from a single trunk of Japanese cedar using recreated stone tools modeled after those from 30,000 years ago. They then undertook a 225-kilometer journey across open ocean, traveling from eastern Taiwan to Yonaguni Island, part of the Ryukyu archipelago (which includes Okinawa), using only natural cues such as the sun, stars, ocean swells, and their own intuition for navigation.

The voyage lasted more than 45 hours, much of it spent without any visual contact with their destination. Even years later, the team continues to analyze the data collected during the expedition, using their findings to refine and validate models of how ancient humans may have completed similar sea crossings.

“A dugout canoe was our last candidate among the possible Paleolithic seagoing crafts for the region. We first hypothesized that Paleolithic people used rafts, but after a series of experiments, we learned that these rafts are too slow to cross the Kuroshio and are not durable enough,” said Kaifu. “We now know that these canoes are fast and durable enough to make the crossing, but that’s only half the story. Those male and female pioneers must have all been experienced paddlers with effective strategies and a strong will to explore the unknown. We do not think a return journey was possible. If you have a map and know the flow pattern of the Kuroshio, you can plan a return journey, but such things probably did not take place until much later in history.”

Ocean modeling and virtual voyages

To explore how varying conditions might have affected the success of such a journey, the researchers turned to sophisticated ocean modeling techniques. They ran hundreds of virtual simulations, adjusting variables such as departure locations, seasonal timing, and paddling methods. These tests were conducted using both present-day and reconstructed ancient ocean conditions to evaluate a wide range of possible scenarios.

“I major in oceanography and use numerical methods and particle tracking techniques to research things like eel and salmon migrations, pumice drift after volcanic eruptions, and oil spills in the Gulf of Mexico,” said Yu-Lin Chang from the Japan Agency for Marine-Earth Science and Technology, and a visiting researcher at UTokyo and lead author of one of the papers in this study. “The Kuroshio Current is generally considered dangerous to navigate. I thought if you entered it, you could only drift aimlessly. But the results of our simulations went far beyond what I had imagined. I’m pleased this work helped illuminate how ocean voyages may have occurred 30,000 years ago.”

Strategic knowledge of early seafarers

The simulations helped fill gaps that a one-time experiment could not. They revealed that launching from northern Taiwan offered a better chance of success than from further south, and that paddling slightly southeast rather than directly at the destination was essential for compensating against the powerful current. These findings suggest a high level of strategic seafaring knowledge among early modern humans.

“Scientists try to reconstruct the processes of past human migrations, but it is often difficult to examine how challenging they really were. One important message from the whole project was that our Paleolithic ancestors were real challengers. Like us today, they had to undertake strategic challenges to advance,” said Kaifu. “For example, the ancient Polynesian people had no maps, but they could travel almost the entire Pacific. There are a variety of signs on the ocean to know the right direction, such as visible land masses, heavenly bodies, swells, and winds. We learned parts of such techniques ourselves along the way.”

References: “Traversing the Kuroshio: Paleolithic migration across one of the world’s strongest ocean currents” by Yu-Lin K. Chang, Yasumasa Miyazawa, Xinyu Guo, Sergey Varlamov, Haiyan Yang and Yousuke Kaifu, 25 June 2025, Science Advances.
DOI: 10.1126/sciadv.adv5508

Funding: This work was financially supported by JSPS KAKENHI grants 23K03503, 21H04368 and 22H00027.

“Paleolithic seafaring in East Asia: An experimental test of the dugout canoe hypothesis” by Yousuke Kaifu, Chih-Hsing Lin, Nobuyuki Ikeya, Masahisa Yamada, Akira Iwase, Yu-Lin K. Chang, Masahiro Uchida, Koji Hara, Kunihiro Amemiya, Yunkai Sung, Katsuaki Suzuki, Minoru Muramatsu, Michiko Tanaka, Sayaka Hanai, Toiora Hawira, Saki Uchida, Masaki Fujita, Yasumasa Miyazawa, Kumino Nakamura, Pi-Ling Wen and Akira Goto, 25 June 2025, Science Advances.
DOI: 10.1126/sciadv.adv5507

Funding: This work was financially supported by JSPS KAKENHI grant JP18H03596.

Never miss a breakthrough: Join the SciTechDaily newsletter.