Abstract:Our study employs acoustic monitoring to unravel the presence of anurans and birds across tropical biomes. Passive acoustic monitoring (PAM) is increasingly popular in ecological research, but recording and analyzing large amounts of data is still a critical bottleneck for the long‐term monitoring of multiple species. We evaluated how temporal and spatial sampling effort affects species diversity estimates using a set of 14,045 1‐min recordings from various neotropical birds and anuran communities. Our goals were to evaluate (i) the daily vocal activity cycle of birds and anurans, (ii) the effect of temporal structure (e.g., number of minutes listened each hour; continuous versus intermittent recordings) on determining the species composition, and (iii) the species–area relationship, and how the number of recorders affects species richness estimates. Based on sampling coverage and completeness, we (iv) evaluate manual inspection schedules for birds and anurans across four biomes of Brazil. We found marked diel variation in vocal activity between taxonomic groups, indicating that birds and anurans are more efficiently detected during early periods of the day and night, respectively. For proper diversity estimates, biomes with higher biodiversity required longer inspecting periods and a larger number of replicates, irrespective of taxa. Although fewer recordings per hour are less informative than full‐hour sampling, species diversity is better estimated when inspected minutes are interspersed over longer periods than inspecting minutes recorded over shorter timespans. Based on our findings, we recommend how to set PAM programs over highly diverse ecosystems. in Portuguese is available with online material. Resumo O monitoramento acústico passivo (MAP) vem se tornando uma ferramenta cada vez mais popular na ecologia, mas a gravação e análise de grandes volumes de dados ainda representam um desafio para monitoramentos de longo prazo ou de múltiplas espécies. Avaliamos aqui como o esforço de coleta afeta as estimativas de diversidade usando um conjunto de 14.045 gravações de um minuto contendo aves e anfíbios. Nossos objetivos foram avaliar i) o ciclo diário de atividade vocal, ii) o efeito da estrutura temporal (por exemplo, número de minutos ouvidos a cada hora; gravações contínuas versus intermitentes) na determinação da composição de espécies e iii) utilizar a relação espécie‐área para avaliar como o número de gravadores afeta as estimativas. Com base na cobertura e completude dos dados, iv) avaliamos horários de inspeção manual para aves e anfíbios em quatro biomas do Brasil. Encontramos variação diária marcante na atividade vocal, indicando que aves e anfíbios são detectados de maneira mais eficiente durante os períodos iniciais do dia e da noite, respectivamente. Para estimativas adequadas de diversidade, biomas com maior biodiversidade exigiram períodos de inspeção mais longos e um maior número de repetições, independentemente do táxon. A diversidade é estimada de forma mais eficiente quando os minutos inspecionados são intercalados ao longo de períodos mais extensos do que minutos registrados ao longo de intervalos de tempo mais curtos. Com base em nossas descobertas, fazemos algumas recomendações de como estabelecer programas de PAM em ecossistemas altamente diversos.

soundClass: An automatic sound classification tool for biodiversity monitoring using machine learning

Acoustic animal identification using unsupervised learning

Learning to detect an animal sound from five examples

A machine vision system for avian song classification with CNN’s

AI-Enhanced Acoustic Analysis for Comprehensive Biodiversity Monitoring and Assessment

The shape of sound: A new R package that crosses the bridge between Bioacoustics and Geometric Morphometrics

Multi-Index Ecoacoustics Analysis for Terrestrial Soundscapes: A New Semi-Automated Approach Using Time-Series Motif Discovery and Random Forest Classification

Automated Bioacoustic Monitoring for South African Bird Species on Unlabeled Data

Soundbay: Deep Learning Framework for Marine Mammals and Bioacoustic Research

Bioacoustic classification of avian calls from raw sound waveforms with an open-source deep learning architecture

Wavelet Filters for Automated Recognition of Birdsong in Long‐time Field Recordings

Interpretable and Robust Machine Learning for Exploring and Classifying Soundscape Data

Acoustic Signatures of the Cerrado: Machine Learning Reveals Unique Soundscapes Across Diverse Phytogeographies

Acoustic features as a tool to visualize and explore marine soundscapes: Applications illustrated using marine mammal passive acoustic monitoring datasets

ANIMAL-SPOT enables animal-independent signal detection and classification using deep learning

Automated detection of dolphin whistles with convolutional networks and transfer learning

Applications of machine learning to identify and characterize the sounds produced by fish

Development of deep neural networks for marine mammal call detection using an open-source, user friendly tool

Assessing the potential of BirdNET to infer European bird communities from large-scale ecoacoustic data

Automatic standardized processing and identification of tropical bat calls using deep learning approaches

Acoustic monitoring of anurans and birds in tropical biomes