Open Ocean Reorientation and Challenges of Island Finding by Sea Turtles during Long-Distance Migration

In 1873, Charles Darwin marveled at the ability of sea turtles to find isolated island breeding sites [1], but the details of how sea turtles and other taxa navigate during these migrations remains an open question [2]. Exploring this question using free-living individuals is difficult because, despite thousands of sea turtles being satellite tracked across hundreds of studies [3], most are tracked to mainland coasts where the navigational challenges are easiest. We overcame this problem by recording unique tracks of green turtles (Chelonia mydas) migrating long distances in the Indian Ocean to small oceanic islands. Our work provides some of the best evidence to date, from naturally migrating sea turtles, for an ability to reorient in the open ocean, but only at a crude level. Using individual-based models that incorporated ocean currents, we compared actual migration tracks against candidate navigational models to show that turtles do not reorient at fine scales (e.g., daily), but rather can travel several 100 km off the direct routes to their goal before reorienting, often in the open ocean. Frequently, turtles did not home to small islands with pinpoint accuracy, but rather overshot and/or searched for the target in the final stages of migration. These results from naturally migrating individuals support the suggestion from previous laboratory work [4-6] that turtles use a true navigation system in the open ocean, but their map sense is coarse scale.