Going west: Range expansion for loggerhead sea turtles in the Mediterranean Sea under climate change

Global changes represent possibly the greatest threat to the future of biodiversity, and this is especially true for species using very different habitats during their life cycle. The problem is even greater when dealing with human dominated landscapes (e.g., the Mediterranean basin) where climate change and habitat destruction and degradation often interact synergistically. We explored this synergy focusing on loggerhead sea turtles (Caretta caretta) in the Mediterranean Sea. Sea turtles depend on both marine and terrestrial environments and are influenced at the same time by changes occurring in both realms. To explore the changes of nesting grounds in the last decades, we first analysed the changes in the 10-year geographical centers of gravity for all nests from 1960 to 2020. By focusing only of the last 20 years, we incorporated both terrestrial and marine variables into a species distribution model (SDM), while accounting for temporal variability by using a multi-temporal calibration approach. The center of gravity for all nesting grounds shifted roughly of 1300 km to the northwest, and the climate suitability model (with the lowest AICc value and a mean AUC = 0.919 ± 0.047; p-value ≤ 0.001) highlighted a sharp increase over time in the northwest Mediterranean Sea. In the same time frame, the southeast Mediterranean showed a very limited increase in climate suitability for nesting. The most important variables were anthropogenic variables, which negatively influence nesting probability, and sea surface temperature, with an increase up to a maximum probability of nesting around 24–25 °C, but a rapid decrease at higher temperatures. The potential importance of the North-western Mediterranean beaches as possible nesting range for sea turtles highlights the relevance of proactive efforts to assist sea turtles’ conservation during their range expansion. More in general, our analyses demonstrate the importance of considering variables from multiple realms when modeling the distribution of species with complex life cycles