Year-round carryover effects are driven by migration phenology for Hirundo rustica (Barn Swallow) wintering in West Africa

Recently, population declines have been reported for many migratory birds. Because of complex life cycles, determining the causes for such declines is often difficult. Thus, migratory birds are of special conservation interest. We studied the migratory behavior of barn swallows Hirundo rustica tagged with solar geolocators and determined carryover effects during the entire annual cycle from one breeding season to the next. We used a Partial Least Square Path Model (PLS-PM) to disentangle migratory and breeding events that occur in chronological order. In addition, we controlled for broad environmental conditions in the wintering grounds (NDVI and latitude) and the specific moulting habitat (δ13C). We did not find a carryover effect from reproduction investment in the attachment year to breeding success in the subsequent year. Individuals which invested more in reproduction departed earlier from the breeding colonies, but this in turn did not affect the onset of autumn migration. Thus, the pre-migratory period should be acting as a buffer stage counteracting any previous carryover effects from reproduction investment. On the other hand, we found a long-lasting domino effect from the onset of autumn migration to subsequent breeding success, consistent with the notion of a migratory race. Specifically, individuals which started earlier the autumn migration, arrived earlier to the wintering grounds, started earlier the spring migration, arrived earlier to the breeding colonies and had a higher breeding success. We highlight that the pre-migratory period (i.e., the time elapsed between departure from breeding areas and the onset of autumn migration) should be important for the lifecycle of migratory species, but it has been frequently overlooked.

Methods
This data corresponded to 35 individuals wintering in West Africa (22 females and 13 males) and breeding in southern Spain (Seville and Badajoz). Key dates in migration behavior were obtained after reconstructing the migratory route for each individual.