Potential exposure of loggerhead turtles to pelagic longliners constrains dynamic spatial avoidance in the western Indian Ocean

Fisheries bycatch remains the leading threat to loggerhead turtle (Caretta caretta) populations worldwide, yet interaction risk in the Indian Ocean has been poorly quantified. Here, we assess potential exposure by quantifying the spatiotemporal overlap between oceanic-stage loggerhead turtle foraging habitats and pelagic longline fisheries across the western Indian Ocean. Using satellite telemetry data from tracked individuals exhibiting foraging behaviour, we developed high-resolution daily predictions of foraging habitat suitability by coupling environmental predictors with an ensemble of machine-learning species distribution models. We generated predictions from June 2017 to May 2020 and evaluated them within their area of applicability. We then quantified horizontal overlap with pelagic longline effort derived from Global Fishing Watch and examined the spatial distribution of fishing activity relative to core turtle foraging areas, to assess whether dynamic spatial avoidance based solely on horizontal overlap is feasible. Loggerhead foraging habitats showed marked regional and seasonal contrasts. In the northwest Indian Ocean, suitable habitats were persistent but shifted latitudinally in response to monsoon-driven oceanographic variability, whereas in the southwest Indian Ocean, foraging habitat suitability exhibited strong seasonality associated with the Agulhas Current system and subtropical frontal zones. Overlap with pelagic longlining was substantially higher in the southwest (52% of fishing hours within core foraging areas) than in the northwest (23%), with pronounced seasonal peaks between austral autumn and winter. In both regions, a large proportion of fishing effort occurred well inside core foraging habitats, indicating that effective spatial avoidance would require displacing fishing activity far beyond habitat boundaries. Dynamic spatial management has greater potential to reduce turtle-fishery overlap than permanent static closures, while also outperforming spatially fixed seasonal closures. Nevertheless, our results highlight fundamental limitations of real-time horizontal spatial avoidance measures where fisheries and turtles respond to the same dynamic oceanographic features. While dynamic ocean management may reduce interactions at habitat margins, mitigating loggerhead bycatch in the western Indian Ocean will likely require complementary strategies, including gear-based mitigation, adaptive effort controls, and coordinated regional governance. These findings underscore the need to integrate ecological dynamics and fisheries behaviour, while also accounting for climate variability, when designing bycatch mitigation strategies for highly mobile marine species.