Co-migration fidelity at a stopover site increases over time in African-European migratory landbirds

Abstract

Migratory species are changing their timing of departure from wintering areas and arrival to breeding sites (i.e. migration phenology) in response to climate change to exploit maximum food availability at higher latitudes and improve their fitness. Despite the impact of changing migration phenology at population and community level, the extent to which individual and species-specific response affects associations among co-migrating species has been seldom explored. By applying temporal co-occurrence network models on 15 years of standardized bird ringing data at a spring stopover site, we show that African-European migratory landbirds tend to migrate in well-defined groups of species with high temporal overlap. Such 'co-migration fidelity' significantly increased over the years and was higher in long-distance (trans-Saharan) than in short-distance (North African) migrants. Our findings suggest non-random patterns of associations in co-migrating species, possibly related to the existence of regulatory mechanisms associated with changing climate conditions and different uses of stopover sites, ultimately influencing the global economy of migration of landbirds in the Palearctic-African migration system.

Keywords: avian assemblages; co-occurrence networks; interspecific interactions; migration phenology; stopover.

Figure 1.

Geographical framework of the study area within the Palearctic–African migration system. The dotted circle in the box shows the Pontine islands system with the exact location of Ponza. Colours in the map show a synthetic view of the main African ecoregions corresponding to the supposed wintering areas of species analysed in this study.

Figure 2.

Trend of modularity (Q), showing the degree to which temporal co-occurrence networks subdivide in groups of species characterized by common migration timing. The higher the value, the greater the tendency of networks to cluster into subgroups of species having high temporal co-occurrence. The dashed black line indicates the original time series, while the solid line and the shaded grey area show the measured trends and uncertainty in trend slopes, respectively.

Figure 3.

(a) Heatmap showing the species' co-migration fidelity (i.e. the frequency with which species co-occurred in the same module over the years), with colours corresponding to the measured normalized mutual information, nmi (see the legend at the bottom of the heatmap). Dendrograms above and alongside the heatmap represent the cluster subdivision of species based on the nmi values, using the farthest neighbour method to calculate distance between clusters in hierarchical clustering. Coloured boxes below and alongside dendrograms show to which foraging niche and wintering area, respectively, a species belongs (see the legend at the bottom of the heatmap). Thick-edged squares within the heatmap show the main cluster subdivision of species obtained by cutting the dendrogram at a pre-defined height of 0.5, corresponding to a minimum level of similarity of 50%. (b) Boxplot showing the nmi distribution in species having different wintering areas and foraging niches. Black lines alongside boxplots indicate the significance levels (ns, not significant; *p < 0.05). Two foraging niche categories (granivore ground and omnivorous arboreal) were not included in the analysis due to the presence of only one species each (Streptopelia turtur and Oriolus oriolus).