Disease outbreaks can have drastic impacts on wild animal populations, particularly in colonial animals such as seals and seabirds that congregate in high densities at breeding sites. For example, phocine distemper outbreaks resulted in mass mortality in the Northern European harbour seal (Phoca vitulina) population in 1988 and 2002, respectively 1. In 2021, High Pathogenicity Avian influenza (HPAI H5Nx) spilled over into European seabirds for the first time and by 2022 the outbreak had caused mass mortality on a global scale, particularly in great skuas (Stercocarius skua), sandwich terns (Thalasseus sandvicensis)2,3 and Northern gannets (gannet hereafter, Morus bassanus)4, however the mechanisms underpinning the observed rapid spread of this novel threat to seabirds are unknown.
Mass mortality severely disturbs the breeding environment of surviving animals, but the behavioural, ecological and epidemiological consequences of disease-driven disturbance are poorly understood, since monitoring fine-scale animal behaviour and movement rarely coincides with such unpredictable events. However, animals can modify their behaviour and movements in response to anthropogenic disturbances 5, as evidenced by changes in space use patterns or shifts to nocturnality in response to human disturbances 6,7. Such modifications can also have pronounced consequences for the transmission of diseases. For example, culling, as an effort to control disease, has instead promoted pathogen spatial spread in vampire bats, Desmodus rotundus, and badgers, Meles meles, since the disturbance associated with culling increased dispersal 8,9. However, questions remain over the possibility for positive feedbacks between infection, mobility and subsequent increased transmissivity.
Here, we investigated the impact of an HPAIV outbreak on the movement behaviour of adult gannets before, during and shortly after the epidemic at Bass Rock, UK, the world’s largest gannet colony with ~ 75,000 breeding pairs 10. The movement ecology of adult breeding gannets is exceptionally well studied with satellite and GPS tracking studies from 1998 to the present 11–24, providing a data-rich baseline for comparison. Adult gannets are highly site-faithful, returning each year to the same colony and, mostly, the same nest site 25. If a breeding attempt fails, birds continue nest site ownership14 and maintain a regular routine of commuting between colony and foraging sites at sea 14. Incubating and chick-rearing birds display highly predictable movements to individual-specific foraging areas 14,24 and forage in colony-specific, non-overlapping ranges during breeding 11. Taken together, these behavioural and space use characteristics suggest limited potential for contact among breeding colonies, making adult gannets unlikely candidates for disease transmission across the metapopulation.
We deployed 18 g nanofix GPS-GSM devices (Pathtrack Ltd.) with TESA tape to the tail of ten gannets from five breeding pairs in April 2022, before egg-laying, and confirmed that each pair incubated an egg in late May 2022. On the 4 June 2022, the first clinical symptoms of HPAIV were observed in gannets on the Bass Rock and an outbreak of clade 3.2.4.4b HPAIV H5N1 was confirmed shortly thereafter4. Over the following month, the outbreak reduced the colony size by 72 % and suppressed adult apparent survival to 0.455 %4. All study birds had lost their egg/chick by 15 June 2022, six birds survived the outbreak and one bird was found dead (Figure 1D, Table S1).
The birds which kept transmitting during the HPAI outbreak instigated unprecedented long-distance movements away from the Bass Rock (Figure 1 A,B,C), significantly differing from movement patterns of 120 incubating and chick rearing gannets GPS tagged in 2015 - 2019 at Bass Rock (Figure 1 B, E). In 2015-2019, maximum travel distances from the colony peaked in late May during incubation before decreasing and levelling off as chick-rearing progressed (Figure 1C, E, Julian day: non-HPAIV year, edf = 4.42, F = 12.62, p < 0.01, mixed effect generalised additive model). In contrast, during the HPAIV outbreak in 2022, maximum travel distances increased throughout June, peaking in mid-July before decreasing around 2.5 months after the onset of the outbreak (Figure 1E, Julian day:HPAIV year, edf = 8.21, F = 30.41, p < 0.01).
During these unusual long-distance movements, two GPS-tracked gannets visited other gannet breeding colonies (Figure 1 B, SM1 Figure S1 – S5) and a third bird visited freshwater lakes ~ 50 km inland in Norway before transmission ceased (Figure 1 B, SM1 Figure S6 - S7). None of the 120 birds tracked on the Bass Rock before 2022, nor any of the hundreds of breeding gannets tracked from 16 colonies over more than two decades11,14–16,21,23,24,26,27 , nor any tracked failed breeder 14 has been recorded visiting a different gannet breeding colony or inland area. Prospecting different colonies provides colonial animals with information on breeding site quality, which can inform recruitment 28, but in seabirds, prospecting is primarily associated with immatures29 or Laridae which tend to have lower philopatry than other seabird taxa 30. In contrast, breeding gannets are highly faithful to their breeding colony, acquiring their breeding sites during a prolonged, high-investment process over several years 25. The breakdown of that site fidelity shown here highlights the severity of the disturbance generated by the HPAI outbreak. It appears to be a short-term response, albeit at the peak infectious period, since gannets with surviving chicks GPS-tracked several months after the disease outbreak maintained their well-established routines31.
The study animals were highly likely exposed to the virus since they nested in an area where in-situ observations showed severe mortality of adult breeders4. The duration of viral shedding for HPAIV for seabirds is unknown, but data from six experimentally infected wild duck species suggests that live virus can be shed from 1-2 days after inoculation, for periods between 5 and 14 days 32, hence the onset and duration of long-distance movements described here fall well within such a timeframe.
The proportion of survivors in our small sample that performed unprecedented long-distance movements is astonishingly high. If the pattern scales up to the colony level, the outbreak of HPAIV may self-propagate transmission by altering gannet movement and aggravating the connectivity between different gannet colonies as well as causing incursions into novel habitat types with susceptible animal communities. Such super-spreading 33, even by a small number of individuals, might explain the rapid spread through the gannet colony networks in the West and East Atlantic4.