Acoustic activity monitoring of the Magellanic Woodpecker, an indicator species from the temperate forests of South America

Woodpeckers are a bird group having a key role in ecosystems and which also serve as indicators of habitat quality. Woodpeckers are easily detected by characteristic sounds that comprise vocalisations and drumming. The analyses of massive data resulting from passive acoustic monitoring (PAM) requires sound automated recognition methods, among which BirdNET is becoming the most used algorithm. The Magellanic Woodpecker is an endemic species from the temperate forests in Chile and Argentina. It is the largest woodpecker from South America and its role as an indicator of forest condition and also as an ecosystem engineer has been well established. The current study has two main objectives: 1) to describe the acoustic activity patterns of the Magellanic Woodpecker through PAM and 2) to evaluate the use of BirdNET for the automated analysis of data resulting from PAM on this species. Our data indicate that the sounds produced by the Magellanic Woodpecker have different temporal patterns and our analyses suggest that a manual processing of PAM data is needed for a preliminary characterisation of the sound activity patterns. Our study indicates that training BirdNET with local data enhances the recognition of Magellanic Woodpecker sounds.