This Is How How AI Is Helping Monitor And Count Puffins

The Isle of May is a 140-acre Nature Reserve off the coast of Scotland. Thousands of Atlantic Puffins return to their breeding colony on the Isle of May each year. In March 2020, there were around 80,000 puffins recorded on the island, making it the third-largest colony in the UK.

Puffins make the long journey from northern Africa through Spain’s Canary Islands to several Scottish islands, including the Isle of May, Fair Isle, Lunga and Noss National Nature Reserve. Scotland is home to 80% of the British and Irish puffin populations.

In early February 2023, The Herald reported that Spanish ornithologists and environmental organizations cited nine dead specimens in San Sebastian, and 140 dead puffins were recovered from beaches along the Tenerife coastline. In the news article, the Spanish Ornithological Society confirmed to The Herald puffins discovered on the Spanish coast with leg rings, and more of them had been ringed in Scotland.

Understanding and counting puffin populations are critical to understanding how they are impacted by changing weather as a result of climate change but also by the development of wind farms around colonies.

On the Isle of May, SSE Renewables’ sustainability and wildlife project is using artificial intelligence (AI), machine learning and image recognition technology to examine how puffins are monitored and counted. The five year project, supported by Microsoft, Avanade and NatureScot, hopes to use this data and intelligence to understand better how nearby wind farms could impact a colony of puffins.

According to Michael Wignall, Azure Business Lead, Microsoft UK, Technology, traditionally counting puffins has been a tricky task, with rangers lying on the ground, extending their hands into burrows to feel for a pair of puffins and their egg – leading to a not too hostile response from the birds.

“While puffins are not at imminent risk of extinction, they are on the Birds of Conservation Concern 4 Red List, meaning that there are serious concerns over numbers in the wild,” said Wignall. “Understanding the impact on local wildlife when deploying new building projects, such as offshore wind farms, is vital and helping keep track of the birds’ population flux.”

Wignall says that utilizing AI technology is minimizing disruption to the birds’ breeding and feeding habits. The technology can also detect and differentiate individual puffins from their colony.

“The AI tools recognize each seabird individually, monitoring their movements and activity,” said Wignall. “With around 80,000 puffins recorded in March 2020, this individual understanding of the puffins within the colony would not have been possible without deploying AI technology.”

Wignall says the technology has provided an understanding of the Puffin’s return to land to breed in late March and early April following their eight months at sea and how the birds return to the same burrows as in previous years.

To spot, recognize and count puffins on the Isle of May, the project team is four cameras in stainless steel boxes around the Island that capture live footage of the puffins. The cameras for the initial trial were installed in April 2021, with monitoring taking place over the summer breeding, April to August 2021.

“Each box has a condensation heater, wipers to cope with the weather and a backup power supply,” said Wignall. “Data captured is being stored in Microsoft Azure Data Lake and uses the Azure Kubernetes Service, which has the power and elasticity to handle huge amounts of information.”

When the cameras are on, the AI can spot the puffins, separate them from background images such as rocks and track them individually, frame by frame, as they move around.

Wignall says that while the Isle of May project is focused on puffins, the technology has the potential to monitor other animals who may be affected by developments and infrastructure changes, such as the deployment of new or existing wind farms.

“The technology can work in any environment in which you want to be able to monitor a species and be hands-off, either because it’s too remote or because you don’t want humans interfering in that environment,” said Wignall. “Monitoring salmon, for example, to make sure they can migrate in rivers, is another example of a wildlife colony where AI could be deployed.”

“Puffins are wonderful, and they are a great start, but actually it’s endless – dolphins to porpoises and the birdlife,” said Rachel McEwen, Chief Sustainability Officer at SSE. “In order to really monitor the impact that you’re having and then, of course, that allows you to adapt what you’re doing to reduce the impact.”

The project on the Isle of May begins with the 2023 breeding season.