gui yong nian - stock.adobe.com
Network Rail has used Google’s artificial intelligence (AI) to power a study, conducted in collaboration with London Zoo (ZSL) and Google, to monitor British species such as birds, bats, foxes and hedgehogs.
The UK has lost nearly half of its biodiversity to urban spread and agriculture since the industrial revolution, ranking in the bottom 10% of nations globally.
Network Rail believes the railway network has the potential to support the conservation of British wildlife. Its 2020 Biodiversity Action Plan is a public commitment by the rail network operator to improve lineside biodiversity, including achieving no net loss in biodiversity by 2024, biodiversity net gain by 2040, and maximising the value and connectivity of their landholdings as wildlife corridors.
Google’s AI is being used in an initial study of the Greater London area to improve the speed of analysing tens of thousands of hours of data files and audio, captured by acoustic monitors along the South London network
The preliminary study also aims to identify habitats and uncover indicator species such as the Eurasian Blackcap, Blackbird and Great Tit in some parts of South London, which, according to Network Rail and ZSL, indicates a healthy environment and an important biodiversity benchmark.
Later this year, the two organisations plan to build out the research to new locations and for more species, such as invertebrates and small mammals.
In a blog describing the study, Anthony Dancer, monitoring and technology lead for conservation and policy at ZSL, and Omer Mahmood, executive sponsor for ZSL Partnership at Google Cloud, discussed the approach Google, ZSL and Network Rail have taken, and discussed how British wildlife was declining at unprecedented rates,” wrote Dancer and Mahmood. “The UK has lost nearly half of its biodiversity to urban spread and agriculture since the industrial revolution, ranking in the bottom 10% of nations globally and worst among G7 countries. These declines continue today: the 2019 UK state of nature report found a further 13% decrease in the average abundance of species since scientific monitoring began in the 1970s.”
Read more about AI monitoring
Network Rail is one of the country’s largest public landowners, with a total estate of approximately 52,000 hectares across its 20,000 kilometres of railway corridor. The vast railway network operated by Network Rail includes extensive tracts of green space, including verges, unused tracks and other lineside landholdings. These landholdings traverse all major British terrestrial habitat types, and are home to a rich array of rare, protected and valued species.
Dancer and Mahmood said the railways could act as refuges for wildlife in urban or intensive agricultural landscapes, as well as connecting fragmented habitats and enabling species movement via corridors, or even acting as barriers to species dispersal where tracks divide habitats. “With careful monitoring and management, the rail network could help protect and restore our wildlife, for the benefit of nature and people,” they said.
During the preliminary study, Network Rail and ZSL deployed technology for remote and automated monitoring of wildlife and anthropogenic activity. These included networked and standalone cameras, and acoustic sensors. The study evaluated the viability of using these sensors to assess Network Rail’s progress towards their biodiversity mission.
Vertex AI and Looker Studio were used to identify and map species to the Network Rail estate. ZSL captured 3,000 hours of audio, recorded by 33 acoustic monitors placed across Network Rail’s estate in South London during 2022. Working with Google Cloud, the ZSL team was able to identify “lineside habitats” using this data.
By using the pre-trained machine learning models BirdNet, BatDetect, and CityNet, Google, Network Rail and ZSL said the AI system was able to detect birds, bats and anthropogenic sounds respectively. Using a combination of these models, they were able to identify types of species and map the occurrences to a geographical location.
The data collected was pushed into Cloud Storage File System in User Space (Fuse), which allowed the team to access the audio data files. Vertex AI was used for rapid prototyping and testing machine learning pipelines.
Once the predictions for each model were run on all Network Rail audio recordings, the data was further transformed in BigQuery to calculate the frequency of each species for each geographic location and other trends.
The predictions from each model were combined to build a single prediction for each species, which was then transformed into a frequency count, which the team used to calculate the relative abundance of each species at a given location.
The final stage involved using Looker Studio to map the biodiversity denoted for each species by volume on a map.
ZSL said it plans to use the methods and tools developed during this project to help Network Rail monitor, understand and improve its impact on lineside biodiversity, with a focus on technologies which allow safe, rapid and remote monitoring.