AI enables massive survey of British woodland bats

A massive three-year study across 310 woodlands has concluded Britain's largest woodland bat survey to date. Scientists deployed recording equipment mounted on poles throughout tree canopies across the nation, collecting a total of 90,000 hours of acoustic data containing 98 million bat calls.

According to researchers, examining such an enormous quantity of audio files manually would have taken more than three decades, making a project of this magnitude practically impossible without automated assistance. Instead, scientists deployed open-source artificial intelligence systems to evaluate the recordings, allowing them to detect national species trends and recognize call patterns across thousands of nights.

Tracking elusive species The initiative recorded some of the rarest and most vulnerable bats in the United Kingdom, including the red-listed Barbastelle. The Barbastelle naturally inhabits ancient woodland environments, but its secretive habits have long made the species exceptionally difficult for researchers to monitor.

Data from the study showed that high concentrations of Barbastelle activity were focused across the South West of England. Major activity hotspots were pinpointed in counties such as Dorset, Cornwall, Devon, and Somerset.

Representatives from the participating organizations noted that Somerset serves as a vital sanctuary for the country's bat populations. The appearance of several highly rare species in Somerset during the survey created significant interest among the conservation team.

The nationwide project was conducted jointly by researchers from University College London, the Bat Conservation Trust, and Forest Research. Project leaders stated that artificial intelligence has dramatically altered the scale at which scientists can track natural ecosystems, offering an unprecedented mechanism for large-scale biodiversity assessment.