Spatial cognition, navigation & migration in seabirds
The Manx Shearwater
Based primarily on Skomer Island (Pembrokeshire, Wales), but also multiple colonies across the breeding range, the Manx shearwater is our main study system to look at these kinds of questions (although we also work on other species such as black-browed albatross and the critically endangered Balearic shearwater, and even homing pigeons). Manx shearwaters, like other Procellariiforms, are some of nature's greatest navigators, commuting between foraging locations and their colony over hundreds and sometimes thousands of kilometres. Additionally, Manx shearwaters are global migrants, spending the boreal winter off Argentina. Using miniaturised datalogging GPS and remote-download GPS/GSM devices, we combine free-ranging tracking, displacement experiments and juvenile tracking to investigate the learning mechanisms and sensory systems that underpin their navigation as adults and how young birds establish migratory routes and re-find their natal colonies when they begin to breed several years later. In a recent project we are also investigating how shearwaters may be affected by the accelerating development of floating offshore wind farms in the Cdeltic seas and elsewhere.
A technological revolution over the past 20 years has resulted in smaller devices which collect higher resolution from a greater variety of data sensors than ever before. In general, the GPS that we use are not custom built for shearwaters but, rather, are originally designed for geo-referencing photographing on backpacking expeditions. We take these devices apart and waterproof them using heatshrink plastic. Recently, more sophisticated devices which are purpose built and collect information on depth and acceleration have become available allowing an even clearer picture to be gained about the at-sea lives of seabirds. In parallel we used archival Geolocator-immersion loggers to plot the daily movement and behaviour of shearwaters throughout migration and across multiple years, to understand how events around the annual cycle (and the globe) are linked and interact with environmental change.
We use an 'ethoinformatics' approach to analysing data from animal-borne sensors. This involes the application of supervised and unsupervised machine learning techniques applied to these datastreams to identify, map and interpret discrete behavioural states. Similar analyses can be used to identify decision points in birds' trajectories also, such as when they begin to feed or when they begin to navigate home to the colony.