• tracking
• exogenous cues
• animal movement

Many vertebrate species embark in short or long migratory movements in order to make the most of the available resources. The animal’s decision on when to start and where to orient its movement is driven by an integration of the internal state of the individual with the external cues perceived: firstly, the migrant has to decide when is the right moment to start the migration in order to find good conditions along the route and at their final destination; and then it has to decide where to move, i.e. it has to orient towards the final goal. Apart from the central role played by the internal status of the animal in starting the migratory behaviour, the importance of external stimuli in influencing the final movement and route followed by the migrant, cannot be underestimated. Indeed, once the migration has started, environmental stimuli continue to have an effect on the animal movement by affecting the speed, accuracy and timing of migration, helping the individual in scheduling its arrival to the final destination (again related to the “when to move” issue) and influencing the actual path followed by the migrant (linked to the “where to move” problem).
With the present thesis, I aimed at evaluating the role of environmental factors on the spatial behaviour of three migratory species: the European teal (Anas crecca), the loggerhead (Caretta caretta) and green sea turtle (Chelonia mydas) that were satellite tracked during their long-distance migrations. Teal data was used to answer the “when to go” question by assessing the influence of external cues on the onset of their spring migration.
Conversely, sea turtle data was used to address the “where to go” issue investigating the turtles’ responses to the effect of sea currents during their migrations.
Due to a general lack of information regarding teal migratory flyways in Europe, I firstly investigated teal migratory phenology by analysing tracking data of individuals wintering in three Italian sites. In this way I determined the course, speed and duration of their migratory movements, together with the number and length of stopovers made along the route. Most of the tracked teals left the wintering grounds between mid-February and March following a straight and direct route along the Black Sea-Mediterranean flyway, and reached the breeding sites located in North-Eastern Europe in May. Along their migratory route most birds stopped for several weeks at stopover sites, especially at the very beginning of migration, and this led to a slow overall migratory speed (36 km/d on average). Nevertheless, the active flight segments of the migration were covered at much higher speeds, up to 872 km/d.
To correctly interpret the birds’ migratory strategy, it is crucial to properly identify the start of migration. Nevertheless, at present a univocal and objective method to assess the start of migration is missing, so I performed a comparative analysis between three different methods to determine the onset of migration using teal tracking data as an example.
The starting dates identified by the tested methods were not always comparable, with the greatest inconsistencies among techniques being evident when the birds started early the migration and then stopped for a long period (>20 days) in a stopover area. This shows that caution is needed when choosing one technique over another, as the method used is likely to strongly affect any following analyses on migratory phenology.
After having defined the migratory phenology of this species, I finally focused on investigating the effect of environmental cues on teal’s decision to start the migration. My analysis aimed at assessing if temperature and winds experienced during teal’s sojourn in an area could have had an effect on the decision to leave their wintering grounds and the site of the longest stopover, i.e. a stopover frequented at the beginning of migration for more than 20 days. I run two distinct Cox proportional models for wintering and the longest stopover area, including photoperiod as a control to determine if the models interpreted correctly the animal choices. As already noted for some goose species, the start of migration from the wintering ground was only triggered by the progressive increase of daylight hours, but, once the migration had started, the photoperiod lost its importance and teals took into consideration only the experienced temperatures to continue their migration from the longest stopover area. Winds experienced before departure, on the other hand, did not seem to play a role in teal’s decision to leave or not an area.
The second part of my work addressed the “where to go” issue, specifically investigating the effect of sea currents on turtles during their open-sea migrations. Given that every swimming animal is affected by the flow of the medium it is moving in, the reconstructed route obtained from satellite-derived positions may not always be representative of the actual movement of the turtle and of its swimming speed. To assess the turtle’s active contribution to the tracked, ground-derived movement, I performed a vector subtraction between the ground-derived and sea currents velocities of three turtle populations migrating in quite different oceanographic conditions. The effect of sea currents along the turtle’s migratory routes was not homogeneous among populations, being greater in turtles migrating in the Indian Ocean than in those moving in the South Atlantic Ocean or in the Mediterranean Sea. These results thus highlight the importance of evaluating the sea currents encountered en-route by migrating turtles when studying their movements and inferring the behaviour, speed, etc. from tracking data. Deriving information on turtle’s migratory behaviour simply from tracking data and considering it representative of the actual active speed of the migrant, can indeed lead to errors, especially if the turtle is encountering strong and variable current fields along its route.
The last aspect I investigated is strictly connected with the previous analysis, where it was observed that turtles nesting at Ascension Island and returning to the Brazilian coast at the end of their breeding season, mostly encountered weak currents flowing in the same direction of the turtle heading. This particular condition made me to hypothesise that this population could rely on a very simple navigational mechanism to reach their foraging grounds, like following a given direction for a certain amount of time (vector navigation).
To test this hypothesis, I simulated the post- and pre-nesting migrations of turtles relying on different navigational strategies and, only for the post nesting migration, I compared the modelled turtle routes with the actual routes of turtles tracked. From the simulations obtained it was evident that the post nesting migrations towards a large target like the continental coast could indeed be performed following a simple vector strategy. The pre nesting migrations towards Ascension, instead, require a more complex system, like true navigation, given the difficulties in reaching such a small target located in the middle of the ocean. Thus, the navigation system of Ascension turtles could be another example of a coexistence of different navigational mechanisms in the same animal, that are used in different migrations to accomplish different orientation tasks.