The incidence of bird sounds, and other categories of non-focal sounds, confound the relationships between acoustic indices and bird species richness in southern China

Global Ecology and Conservation, Год журнала: 2024, Номер 51, С. e02922 - e02922

Опубликована: Апрель 1, 2024

The analysis of audio recordings through acoustic indices has been proposed as an efficient way to measure and monitor biodiversity, given the assumption that higher levels biodiversity produce more rapidly-changing complex sound. However, in previous work south China, we have found only moderate correlations between bird species richness, when analyzing were made at same time conducting point counts birds. Here, extended three study regions Guangxi Province, making observations both inside relatively undisturbed forest reserves surrounding agricultural lands. We 42 richness (seven commonly used [ACI, ADI, AEI, AR, BIO, H, NDSI] each calculated six habitat/region combinations) seven had even a relationship (i.e., |r| ≥ 0.20). To understand this paucity relationships, listened subsample recordings, scoring them for various kinds biophony (incidence insect sounds), anthropophony observer-produced sounds, sounds by other humans, machines) geophony (primarily wind, occasionally water). These analyses insects positively related indices, especially forests. Moreover, geophony, were, general, negatively correlated indices. incidence vocalizations also consistent strong residuals models relating demonstrating are clearly sensitive amount but is not necessarily strongly diversity. Multivariate confirmed amounts different sound categories influential on than richness. Our results demonstrate how competing noises, environments heavily modified ones, can confound relationships biodiversity. discuss possible remedial steps using noisy soundscapes, including tailoring selection frequency bands objectives project, or, alternatively, big-data approach combines multiple

Язык: Английский

---

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

Methods in Ecology and Evolution, Год журнала: 2024, Номер 15(6), С. 1071 - 1083

Опубликована: Май 10, 2024

Abstract Passive acoustic monitoring (PAM) has become an important tool for surveying birds, and there is a growing demand approaches to obtain abundance behavioural information from PAM recordings. Changes in bird populations have been assessed by counting recorded calls calculating the vocal activity rate (VAR, i.e. number of per recording time). However, could be counted various ways depending on species traits, these call counts give us different insights abundance, behaviour and/or habitat use. Our study had two goals: (1) present evaluate new indices based counts, detection (DR, 1‐min recordings which presence target vocalization detected) maximum count minute (MAX, found recording); (2) conceptual framework showing how interpretations VAR, DR MAX depend index traits. For three Neotropical with distinct we calculated data 25 sites Yucatan Peninsula (Mexico) that varied their degree anthropogenic disturbance. We moderate high correlations between higher temporal variability VAR compared MAX. also effect sizes disturbance indices. suggest might more reliable relative than whose calling exhibits cue may suitable estimating family or flock size gregarious birds. findings show potential usefulness developing generate ecological hypotheses assess changes behaviour.

Язык: Английский

---

Mitigating bias in long‐term terrestrial ecoacoustic studies

Journal of Applied Ecology, Год журнала: 2025, Номер unknown

Опубликована: Фев. 14, 2025

Abstract Long‐term biodiversity monitoring is needed to track progress towards ambitious global targets reduce species loss and restore ecosystems. The recent development of cheap robust acoustic recording devices offers a cost‐effective means gathering standardised long‐term datasets. Accounting for sources bias in ecological research fundamental part the study design process. To highlight this issue context terrestrial ecoacoustic monitoring, here we collate discuss arising from (i) hardware devices, (ii) firmware, software analysis tools (iii) deployment environment. One important source unavoidable changes hardware—to demonstrate how potentially introduces bias, present two case studies comparing output simultaneous recordings different recorders. mitigate biases, recommend effective documentation environmental hardware‐related variables, as well data storage strategy that facilitates reanalysis. Additionally, use regular calibration tests measure variation detection space will facilitate analytical approaches or post‐hoc AI solutions remove unwanted biases. Synthesis applications : suggested mitigations described be relevance manufacturers, researchers conservation practitioners. Researchers practitioners must fully aware relevant biases when designing should incorporate appropriate into their design.

Язык: Английский

---

Unsupervised acoustic classification of individual gibbon females and the implications for passive acoustic monitoring

Methods in Ecology and Evolution, Год журнала: 2020, Номер 12(2), С. 328 - 341

Опубликована: Окт. 29, 2020

Abstract Passive acoustic monitoring (PAM) has the potential to greatly improve our ability monitor cryptic yet vocal animals. Advances in automated signal detection have increased scope of PAM, but distinguishing between individuals—which is necessary for density estimation—remains a major challenge. When individual identity known, supervised classification techniques can be used distinguish individuals. Supervised methods require labelled training data, whereas unsupervised do not. If signals individuals are sufficiently different, number clusters might represent sampled. The majority applications animal vocalizations focused on quantifying species‐specific call repertoires. However, with interest PAM applications, that needed. Here we use an existing dataset Bornean gibbon female calls known from five sites Malaysian Borneo test three different clustering algorithms (affinity propagation, K ‐medoids and Gaussian mixture model‐based clustering) Calls females readily distinguishable using techniques. For internal validation cluster solutions, calculated silhouette coefficients. external validation, compared results labels standard metric: normalized mutual information. We also accuracy by assigning solutions based which had highest particular female. found affinity propagation consistently outperformed other all metrics used. In particular, was more consistent as increased, when randomly sampled across sites. conclude may useful providing additional information regarding applications. stress although gibbons case study, these will applicable any individually distinct animal.

Язык: Английский

---

Passive acoustic monitoring in terrestrial vertebrates: a review

Bioacoustics, Год журнала: 2023, Номер 32(5), С. 506 - 531

Опубликована: Май 10, 2023

Passive acoustic monitoring (PAM) has become increasingly popular in ecological studies, but its efficacy for assessing overall terrestrial vertebrate biodiversity is unclear. To quantify this, performance species detection must be directly compared to that obtained using traditional observer-based (OBM). Here, we review such comparisons across all major classes and identify factors impacting PAM performance. From 41 found while PAM-OBM have been made classes, most focused on birds (65%) North America (52%). performed equally well or better (61%) OBM general. We no statistical difference between the methods total number of detected (excluding reptiles); however, recording period region study influenced relative PAM, analysis method which sampled longer showed impact. Further studies comparing non-avian vertebrates standardised are needed investigate more detail may influence While a valuable tool surveys, combined approach with targeted non-vocal should achieve comprehensive assessment communities.

Язык: Английский

---