Dispersal is a critical process influencing population dynamics, and patterns of dispersal movement impact species distribution and abundance and have important evolutionary and ecological consequences. In particular, long-distance dispersal (LDD) can be especially important for gene flow and adaptability, although little is known about the mechanisms of LDD because of the challenges of studying large-scale animal movement. Global change is driving selection for new movement patterns in animals by stressing physiological tolerances and affecting the dynamics of biotic interactions. This is resulting in shifts in species distributions that are widespread across taxa. The distributions of migratory bird species are shifting heterogeneously, and multidirectional shifts are inconsistent with unidirectional climate niche tracking hypotheses. We studied correlates of natal LDD using bird banding data for American kestrels (Falco sparverius) in the United States and Canada from 1960-2015. We used Bayesian hierarchical models to investigate temporal trends and the effects of sex, migration strategy, weather, and landcover on LDD frequency and distance and understand potential sampling bias. We also studied the directions and correlates of shifts in breeding distributions for 73 avian species and subspecies from 1994-2017 using the North American Breeding Bird Survey. We modeled regional changes over time in breeding abundance centroid and investigated the effects of abundance trends and migratory, habitat, and dietary traits on these shifts.
Nearly half of all natal dispersal (48.7%) in kestrels was LDD (> 30 km), and the likelihood of LDD was positively associated with agriculture at natal sites. LDD distance was positively correlated with latitude, a proxy for migration strategy, indicating that migratory individuals disperse farther than residents. For male kestrels, LDD was positively associated with maximum summer temperature. Unlike previous studies of short-distance dispersal (SDD), we did not find female-bias in either LDD frequency or distance. Sampling affected frequency and magnitude of LDD, likely because local studies more frequently capture SDD within study areas.
In our study of breeding distributions, we found that 44% of regional shifts were equatorward, 55% were poleward, and several species shifted in different directions in different regions. We did not find any life history traits that explained southward shifts, but diet, migratory strategy, and tolerance to humans predicted northward shifts. Our results clearly indicate the prevalence of multidirectional breeding distribution shifts, and suggest that life history is one component in a likely complex set of interacting mechanisms acting at many scales to drive shifts.
Our results show that patterns of dispersal and distributions are complex, and shaped by interactions between environmental factors and life history. Our results that LDD frequency and distance are influenced by different intrinsic and proximate environmental factors from SDD suggest LDD and SDD may be distinct processes rather than originating from a single dispersal distribution. While the drivers of equatorward distribution shifts are still unclear, multidirectional shifts do not support the hypothesis that tracking climate warming is the primary driver of shifts, and investigation into drivers of equatorward shifts is necessary for understanding the heterogeneous effects of climate change on distributions. The feedback between dispersal and distributions is a critical piece of species’ response to global change, and it is important we strive to understand their causes and consequences to further develop our concept of adaptation and persistence in the current era.