Simultaneous passive acoustic monitoring uncovers evidence of potentially overlooked temporal variation in an Amazonian bird community

Abstract

Birds are frequently used as indicator taxon in ecology, but traditional surveys are prone to error and largely inadequate in the dense, speciose tropical forests where they are most needed. Passive acoustic monitoring (PAM) has the potential to address many of these pitfalls, providing unique advantages for both long-term monitoring and rapid biodiversity assessment. Here, we evaluate the effectiveness of PAM for characterizing spatiotemporal variation in avian species richness, diversity, vocal activity, and community composition in the Amazon basin global biodiversity hotspot. We employed an array of ten autonomous recording units (ARUs) spanning edge, degraded, and interior forest at a reserve on the deforestation frontier of Madre de Dios, Peru, a region featuring some of the richest, most complex avian assemblages on Earth. Recordings from 21 dawn-hours across three days at seven sites were manually annotated by a single observer. ARUs and manual annotation performed well as an inventory method, especially for midstory and understory species. Sites in edge and degraded habitats featured significantly lower avian species richness, Shannon diversity, and vocal activity levels than in the forest interior, patterns that replicate the findings of prior high-effort avian censuses in the region. We observed significant temporal variation between days; at all featured sites, vocal activity was highest on January 20th and lowest on January 31st. We demonstrated that novel annotation-generated metrics can work as effective proxies for abundance data and per-capita song rate, correlating with diversity indices and efficiently characterizing habitats at a level of detail unobtainable with human observers. Generalist species were significantly overrepresented in the soundscapes of interior forest, relative to their contributed share of species richness, and interior specialists were underrepresented, suggesting that the entire 191ha site is degraded. The between-days temporal variation that we documented, almost certainly overlooked without simultaneous monitoring, may obfuscate or distort the results of traditional surveys. Synthesis & Applications. We propose that PAM should become a cornerstone of biodiversity research. The standardization advantages demonstrated in this study, permanent storage, multi-taxa applications, and potential of automated identification make bioacoustics an ideal methodological avenue for guiding management and policy under rapid global change.

How to count bird calls? Vocal activity indices may provide different insights into bird abundance and behaviour depending on species traits

“…We calculated DR, MAX and VAR for each of the 4 months in our study period (November and December 2021, January and February 2022) for each monitoring site and for each species. Calculating vocal activity indices based on an adequate monitoring period is essential to account for daily variation in vocal activity when aiming to determine relative abundance (Hopping & Klinck, 2021; Pérez‐Granados, Gómez‐Catasús, et al., 2019). For this purpose, we chose 1 month as a conservative time period, based on recommendations from previous studies (Pérez‐Granados & Schuchmann, 2020a: 3–15 days; Pérez‐Granados, Gómez‐Catasús, et al., 2019: minimum 9 days; Abrahams, 2019: 1 month).…”