Combining UAVs and multi-sensor dataloggers to estimate fine-scale sea turtle density at foraging areas: a case study in the central Mediterranean

ABSTRACT: Knowledge of the distribution and density of marine species is key to understanding habitat use and interactions with human activities. Yet such information for sea turtles remains scarce, especially at foraging areas, where low turtle density represents an additional challenge in comparison to turtle aggregations at coastal breeding areas. Aerial surveys with aircraft are an efficient method for collecting these data over broad scales, while more novel unoccupied aerial vehicles (UAVs) are better suited for finer-scale data collection. However, their use is less developed, especially in offshore areas. Here we explored, for the first time in the Mediterranean, the potential of UAV surveys to estimate turtle density (surface and total) at foraging areas and its temporal trend. Between 2017 and 2023, we conducted 427 flights in the Pelagian Islands Archipelago (PIA), Italy, a regionally important foraging area of the loggerhead sea turtle Caretta caretta. To convert from surface to total density, we used data from multisensor biologgers deployed on 22 turtles to calculate the proportion of time turtles are visible from aerial surveys (availability time proportion, ATP). Results show that the mean surface turtle density in the PIA (0.336-0.477 turtles km^-2) is among the highest reported globally for a loggerhead turtle foraging area. Such densities make it possible to assess population trends through periodic UAV surveys, which are less expensive than aircraft surveys and which can minimize the typical biases of aerial surveys (distance sampling, perception, and misidentification). A standardized methodology is needed for meaningful comparisons, including ATP at the visible depth layer vs. surface.