AI helps record 98 million bat calls in largest UK woodland survey

Researchers recorded 98 million bat calls over three years in 310 woodlands across Britain, in what scientists describe as Britain's largest-ever woodland bat survey. The project used AI to analyse the recordings, which the team says would have taken more than 30 years to process manually.

The survey involved University College London, the Bat Conservation Trust and Forest Research. Devices were placed on poles among the trees to capture the calls.

Rare species detected

Some of Britain's rarest and most threatened bats were recorded, including the red-listed Barbastelle. Hotspots of Barbastelle activity were concentrated in south-west England, including Somerset, Devon, Cornwall and Dorset. The species roams ancient forests but is elusive, making it difficult to track.

Kate Jones, from UCL, said AI had 'fundamentally changed the scale' at which wildlife can be monitored. She called the survey a 'nice story of AI for good'. She said that previously, analysing this volume of acoustic data would not have been feasible. Open-source AI tools allowed the team to identify millions of bat calls from thousands of nights of recordings, find patterns and flag species at a national scale, she added.

'Unprecedented' project

Lia Gilmour, of the Bat Conservation Trust, described the project as 'unprecedented'. She said there was 'really good bat activity' in the south-west area of Somerset, which she described as a stronghold for UK bat species. She added it was exciting that some of the rarest species appeared in Somerset as part of the survey.

Gilmour said AI is seen as a powerful tool in biodiversity monitoring. With 90,000 hours of recording and huge volumes of data, she said there was no way the team could have done the work manually.

Source: BBC Science.