Finding food: deciphering the foraging ‘fingerprints’ of Gannets
Chris Pollock
Blog Topics
The comings and goings of seabirds were shrouded in mystery for many years. Where did they go to find food for their chicks? How long were their migrations, and where did they spend the months outside the breeding season?
The first use of remote satellite tracking for seabirds in 1989 revolutionised our understanding of their behaviour. Since then, there has been an explosion in the number of studies that track the movements and migration of a multitude of species.
There is a lot about seabirds that we still don’t know, however. Seabird ecologists like me are still trying to understand why the species they study occupy certain spaces and behave in certain ways. To this end, we are often tempted to grab as many fancy new tracking devices as our budget can afford and head to the nearest coastline in search of some unsuspecting birds.
These devices provide us with lots of information about the location and movements of the birds, which we can then analyse to see if differences in behaviours – such as time spent resting, foraging and travelling – might be due to differences in sex (male or female birds), age (immature or adult birds) or even the breeding colony location.
These methods have given us great insight into seabird behaviour. But more and more studies are highlighting the importance of understanding the movements and behaviours of individual birds. You could think of this as trying to understand a single bird’s personality.
Introducing the foraging ‘fingerprint’
Individual Gannets leave their breeding colony in the same direction on each foraging trip, search for food in the same area on successive trips, and will even return to that spot from year to year.
One prominent example of this is called ‘individual foraging site fidelity’ (IFSF), which describes a pattern of behaviour well documented in Gannets. Tracking devices have revealed that individual Gannets leave their breeding colony in the same direction on each foraging trip, search for food in the same area on successive trips, and will even return to that spot from year to year.
You can think of the foraging patterns which have been revealed by tracking as a Gannet’s foraging ‘fingerprint’ – a pattern unique to that bird.
Why might a Gannet want to return to the same spot again and again?
Well, Gannets breed in large colonies – the largest in the world is located on Bass Rock, Scotland, and is made up of around 150,000 breeding adults! When foraging from such a big colony, it makes sense to remember an area where you have had a successful foraging trip. There will be a lot of competition for fish from other Gannets, so making your trip as efficient as possible will be beneficial to you and the hungry chick you have to feed back at the colony.
The next question that we must ask ourselves is what processes underpin this foraging fingerprint. For example, Gannets might use memory to recall particularly rich foraging sites, or observe other Gannets to see where they feed successfully, and copy them.
To investigate this question with more tracking devices, we would need to track all the Gannets from a colony at once to understand social interactions. This would be incredibly expensive, not to mention the practical impossibility of attaching 150,000 tags to adult Gannets. And as of yet there are no questionnaire techniques to interrogate Gannets about how they repeatedly find their favourite feeding spots.
This is where simulation models come into play.
Simulation models allow us to investigate the underlying mechanical processes in natural systems. When we create a model, we can construct a scenario and then apply different ‘rules’ for how an organism might interact with others and its environment. Then we can see which rules generate the result closest to that which we see in real life, and are therefore likely to reflect the mechanisms underpinning that reality.
Deciphering the fingerprint
This simulation depicts individual Gannet foraging trips, each departing from the Bass Rock colony in the Firth of Forth, Scotland. Outward trips (yellow) include foraging behaviour (red). The birds return to the colony (blue) after foraging.
My quintessential lightbulb moment occurred during my PhD when reading about a tracking study which suggested that Gannets anticipate their prey location. The study stated that “further investigations are necessary to identify the mechanisms involved in seabird resource localisation, such as sensorial abilities, memory effects, public information [another term for social interactions] or a combination of these factors”.
I’d already begun to work with simulation models, and it dawned on me that they would provide the perfect format to do exactly that. By modelling scenarios with different rules for memory use, the impact of social interaction and combinations of these, I could investigate the factors which determined a Gannet’s individual foraging fingerprint – the departure direction and foraging area used in repeated feeding trips.
I could then see which models produce the best ‘fit’ for the fingerprint that has been revealed by tracking, gaining insight into what Gannets think and how they interact with each other.
I began by hypothesising a range of rules which might govern Gannet foraging behaviour. They incorporated the use of memory (e.g. short or long-term), and how Gannets interacted with each other while on foraging trips (e.g. attraction to others foraging, or avoidance of areas with high concentrations of other Gannets).
In total, I constructed 16 different rules or rule combinations based on memory and social interaction, with varying levels of complexity. Then I ran my simulations of Gannet behaviour governed by each of my hypothesised rules or rule combinations, and analysed the results from a few random individuals in each simulation – like sampling from a real-life population. The results were fascinating.
The simulation which produced the most realistic movements of individual Gannets was governed by both memory and social interaction – the most complex rule combination that we created.
The simulation which produced the most realistic movements of individual Gannets was governed by both memory – remembering several different locations to visit – and social interaction – using other Gannets as cues for where food might be located and as indicators of very high competition. This rule combination was the most complex that we created.
I would argue that this is interesting in and of itself, but what are the potential real-world applications for such a study?
Modelling in the real world
With this information about the mechanisms which govern seabird foraging behaviour, we have a more robust baseline for further research. We can create new models founded on this baseline to investigate how changes to the environment might affect our seabirds.
One of the most prominent potential threats to Gannets and other seabirds from UK colonies is the rapid development of offshore wind farms around our coasts. With government plans to more than triple our capacity within the next 10 years, we must do our best to predict the potential impacts of different development plans on the distributions and population sizes of our treasured seabirds – and modelling can help.
‘Collision risk modelling’ is a type of simulation modelling which investigates the risk of seabirds flying into wind turbines. It is often used to inform the Environmental Impact Assessments conducted for potential developments, for example.
More recently developed models are examining how offshore wind farms might disrupt seabird foraging trips. So far, the modelling techniques we have indicate that the cumulative effects of offshore wind farm development are having an impact on our seabirds.
Looking ahead
We now have the potential to simulate these ‘bird brains’ (historically an insult, although I would say one which is rapidly being disproved) as they learn and memorise things and react to competitors in a myriad of ways. Our simulations are representative of reality and of our ever-increasing knowledge of the complexities of these birds. I look forward to applications looking at immediate threats such as wind farms and changing prey distributions to predict how our feathered friends may fare in these challenging times.
This blog is based on the research and findings of Chris's PhD, Modelling breeding season foraging and tracking autumn migrations to fill knowledge gaps in gannet ecology relating to impacts of offshore wind farms (University of Leeds, 2022).
Share this page