Cory’s, Scopoli’s, and Cabo Verde shearwaters non-breeding locations

Aim: in highly mobile species, Migratory Connectivity (MC) has relevant consequences in population dynamics, genetic mixing, conservation, and management. Additionally, in colonially breeding species, the maintenance of the breeding geographical structure during the non-breeding period, i.e., a strong MC, can promote isolation and population divergence, which ultimately can affect the process of lineage sorting. In geographically structured populations, studying the MC and differences in environmental preferences among colonies, populations, or taxa can improve our understanding of the ecological divergence among them.

Location: Atlantic Ocean and Mediterranean Sea.

Methods: we investigated the MC and non-breeding ecological niche of three seabird taxa from the genus Calonectris (n = 805 individuals). Using 1,346 year-round trips from 34 different breeding colonies we assess the level (from taxa to colony) at which MC,  and non-breeding spatial and environmental segregation, emerge.

Results: at a taxon level, we found a clear difference in the non-breeding distributions between Cory’s (C. borealis) and Scopoli’s (C. diomedea) shearwaters, and a clear ecological divergence between Cory’s and Cape Verde (C. edwardsii) shearwaters. At an intermediate aggregation level we found that birds breeding in proximity had similar non-breeding habitat preferences, while birds breeding in very distant colonies (and therefore classified in different populations) had different non-breeding habitat preferences. Furthermore, within each taxon, we found more structure (i.e., stronger MC) and non-breeding divergence at an intermediate aggregation level than at the colony scale, where MC was weak.

Main conclusions: These results suggests that conspecifics from nearby colonies mix in common non-breeding areas, but not with birds from more distant colonies or different taxa. These results support the need for management and conservation strategies that take in account this structure when dealing with migratory species with high connectivity.

Methods
Positions were obtained with Global Location Sensors (GLS; Wilson et al. 1992), which provide one or two positions per day, using daylight to calculate longitude from the time of twilight and latitude from the length of the day (light period). Detailed deployment information for each colony can be found in Morera‐Pujol et al. (2023). Twilight events calculated from the light measurements were visually inspected and corrected when interferences near the twilights were detected. Locations were obtained from the light data using either Intiproc® (Migrate Technology Ltd.) or Biotrack® (Biotrack Ltd.) software, or the GeoLight package in R (Lisovski and Hahn 2012). To eliminate biases due to incorrect latitude estimations during or near equinoxes (Ekstrom 2004) we removed position data from 20 days before and after the equinoxes. In addition, we removed unrealistic positions by applying a quadratic speed filter following McConnell et al. (1992).

We visually inspected tracks to assign a phenological state to each position. Based on the distance between consecutive positions and directionality of movement, we classified periods as “residency” when for at least four days the distance between consecutive positions was short and the movement non-directional (following Dias et al. 2011). The residency periods during the breeding season and around the breeding colonies were classified as “breeding”. The residency periods outside the breeding season were classified as “main non-breeding” (the longest) and “staging” (the rest of them). Movement periods (long distances between consecutive locations and directionality) were classified as “migration”. For our analysis we used only the residency periods during non-breeding.