The effects of climate change on the global migration of birds

Abstract

Migration is one of the most spectacular wonders of the natural world but is increasingly threatened by large-scale environmental change. Migrants are key components of biological systems at higher latitudes, where the magnitude and velocity of climate change is most pronounced. In this thesis, I consider how the distribution of migratory species may change in the future, across the globe, in response to environmental change.

Using global data on avian species distributions, the environmental and biological factors that affect migratory species richness were investigated. The cost of migration distance between breeding and non-breeding areas, and resident species richness were two important determinants of migrant species richness in a given location. Species distribution models (SDMs) were used to relate migratory species occurrence in breeding and non-breeding seasons to contemporary climatic variables. The importance of climatic variables for predicting migrant species’ occupancy depended on season (breeding or non-breeding). Whilst most SDMs focus on individual species breeding ranges, and model within the occupied realm, this can be problematic for migratory species that frequently use multiple realms, especially if trying to project newly suitable areas in future. Here, I developed a novel method to assess climatic suitability for migratory species globally, within species migration flyways. Selecting absence data to condition SDMs for species occurring across multiple realms can be problematic, as regions distant from the range may be suitable but unoccupied. To minimise this issue, I developed a modelling approach that selected absences preferentially from closer to a species distribution, using a distance weighting function; compared to other methods, this improved model performance on withheld test data. Using this approach, I projected the potential future breeding and non-breeding ranges of all global terrestrial long-distance migrants under climate change. Migrant breeding ranges were projected to shift poleward, but non-breeding range shift projections are less consistent in direction.

To date, predictions of migratory changes under future climate scenarios have usually estimated a single distance between breeding and non-breeding range centroids. However, this approach ignores the variation in migratory movements within species. I developed a method to estimate the range of potential migration distances for individual populations, both now and in the future, which agreed well with recovery data for a well-recorded European trans-Saharan migrant. This approach projected longer median migratory distances in the future for many species. This thesis highlights the importance of climate for migratory species, and suggests that the observed general decline in many long-distance migratory species may be exacerbated by ongoing climate change.