Acoustic indexes as predictors of bird richness: Insights from atlantic forest fragments

Bird communities present in remaining fragments of the Atlantic Forest suffer several impacts. Fragments bordering roads suffer from anthropophonies caused by vehicle traffic, affecting the sound communication, the behaviour, and changing communities structure. In this study, we tested the effect of different types of Atlantic Forest fragments on soundscapes and bird community. The sampled points had different degrees of urbanization, categorized in three Zones (Urban close to roads; Rural close to roads; Protected Areas far from roads). 810 minutes of audio recordings were analysed at 9 different points, using 6 Acoustic Indexes (AIs) to compare the soundscapes. Bird richness was quantified from the identification of calls using pattern matching mechanisms together with manual validations for each minute. The 6 AIs indicated significant differences between the soundscapes of different Zones. A total of 77 species were identified (Urban n = 25; Rural n = 50; UC n = 41). Despite having the lowest number of species, the Urban Zone had the highest richness per minute sampled. We found correlations between AI and bird richness, but it was controversial when compared to previous research, and that correlation could be stochastic. AI can help us monitor changes in the environment at low costs, but it is difficult to predict richness or diversity directly using them. We suggest using Autonomous Recording Units and AI to constantly monitor vulnerable forest fragments. Changes in AI patterns can help determine when and where we should invest and use more direct methods for evaluating richness and diversity.