Macro-scale relationship between body mass and timing of bird migration

Abstract

Clarifying migration timing and its link with underlying drivers is fundamental to understanding the evolution of bird migration. However, previous studies have focused mainly on environmental drivers such as the latitudes of seasonal distributions and migration distance, while the effect of intrinsic biological traits remains unclear. Here, we compile a global dataset on the annual cycle of migratory birds obtained by tracking 1531 individuals and 177 populations from 186 species, and investigate how body mass, a key intrinsic biological trait, influenced timings of the annual cycle using Bayesian structural equation models. We find that body mass has a strong direct effect on departure date from non-breeding and breeding sites, and indirect effects on arrival date at breeding and non-breeding sites, mainly through its effects on migration distance and a carry-over effect. Our results suggest that environmental factors strongly affect the timing of spring migration, while body mass affects the timing of both spring and autumn migration. Our study provides a new foundation for future research on the causes of species distribution and movement.

Fig. 1. Distribution of the breeding (red dots) and non-breeding (blue dots) sites of birds included in this study.

The figure shows data from 1531 individuals and 177 populations, obtained from 306 studies of 186 species. Each dot represents the median longitude and latitude for each species in each study.

Fig. 2. Results of the structural equation model analysis on migration timing of 186 species.

Arrows represent the direct effects of variables on the migration timing of birds. Red arrows and values (correlation coefficients) represent positive significant effects (credible intervals excluding zero); blue arrows and values (correlation coefficients) represent negative significant effects (credible intervals excluding zero). The dashed-line arrows represent non‐significant relationships (credible intervals including zero). R²_m, marginal R square, representing only the variance of the fixed effects, R²_c, conditional R square, representing both the fixed and random effects. The absolute value of the non-breeding latitude was used as the non-breeding latitude.

Fig. 3. Direct effect of body mass on spatiotemporal migration patterns of birds in structural equation models.

Direct effects of body mass on the non-breeding latitude (a), where absolute values were used for birds wintering in the Southern Hemisphere, departure date from the non-breeding site (b), departure date from the breeding site (c), and arrival date at non-breeding site (d) of migratory birds. Light bands represent 95% credible intervals. Dots represent each tracked individual or population. The density distribution of each variable is plotted alongside the scatter plots.

Fig. 4. Relationships of the timings of key events in migratory birds with environmental drivers in structural equation models.

The relationships between migration timings in spring and breeding latitudes (a, b) and the relationships between arrival dates and migration distances (c, d). Light bands represent 95% credible intervals. Dots represent each tracked individual or population. The density distribution of each variable is plotted alongside the scatter plots.