Towards the fully automated monitoring of ecological communities

Ecology Letters, Journal Year: 2022, Volume and Issue: 25(12), P. 2753 - 2775

Published: Oct. 20, 2022

Abstract High‐resolution monitoring is fundamental to understand ecosystems dynamics in an era of global change and biodiversity declines. While real‐time automated abiotic components has been possible for some time, biotic components—for example, individual behaviours traits, species abundance distribution—is far more challenging. Recent technological advancements offer potential solutions achieve this through: (i) increasingly affordable high‐throughput recording hardware, which can collect rich multidimensional data, (ii) accessible artificial intelligence approaches, extract ecological knowledge from large datasets. However, automating the facets communities via such technologies primarily achieved at low spatiotemporal resolutions within limited steps workflow. Here, we review existing data processing that enable communities. We then present novel frameworks combine technologies, forming fully pipelines detect, track, classify count multiple species, record behavioural morphological have previously impossible achieve. Based on these rapidly developing illustrate a solution one greatest challenges ecology: ability generate high‐resolution, standardised across complex ecologies.

Language: Английский

---

Towards a multisensor station for automated biodiversity monitoring

Basic and Applied Ecology, Journal Year: 2022, Volume and Issue: 59, P. 105 - 138

Published: Jan. 7, 2022

Rapid changes of the biosphere observed in recent years are caused by both small and large scale drivers, like shifts temperature, transformations land-use, or energy budget systems. While latter processes easily quantifiable, documentation loss biodiversity community structure is more difficult. Changes organismal abundance diversity barely documented. Censuses species usually fragmentary inferred often spatially, temporally ecologically unsatisfactory simple lists for individual study sites. Thus, detrimental global their drivers remain unrevealed. A major impediment to monitoring lack human taxonomic expertise that implicitly required large-scale fine-grained assessments. Another amount personnel associated costs needed cover scales, inaccessibility remote but nonetheless affected areas. To overcome these limitations we propose a network Automated Multisensor stations Monitoring Diversity (AMMODs) pave way new generation assessment centers. This combines cutting-edge technologies with informatics expert systems conserve knowledge. Each AMMOD station autonomous samplers insects, pollen spores, audio recorders vocalizing animals, sensors volatile organic compounds emitted plants (pVOCs) camera traps mammals invertebrates. AMMODs largely self-containing have ability pre-process data (e.g. noise filtering) prior transmission receiver storage, integration analyses. Installation on sites difficult access require sophisticated challenging system design optimum balance between power requirements, bandwidth transmission, service, operation under all environmental conditions years. An important prerequisite automated identification databases DNA barcodes, animal sounds, pVOCs, images used as training identification. thus become key component advance field research policy delivering at an unprecedented spatial temporal resolution.

Language: Английский

---

Acoustic localization of terrestrial wildlife: Current practices and future opportunities

Ecology and Evolution, Journal Year: 2020, Volume and Issue: 10(13), P. 6794 - 6818

Published: June 13, 2020

Abstract Autonomous acoustic recorders are an increasingly popular method for low‐disturbance, large‐scale monitoring of sound‐producing animals, such as birds, anurans, bats, and other mammals. A specialized use autonomous recording units (ARUs) is localization, in which a vocalizing animal located spatially, usually by quantifying the time delay arrival its sound at array time‐synchronized microphones. To describe trends literature, identify considerations field biologists who wish to these systems, suggest advancements that will improve we comprehensively review published applications wildlife localization terrestrial environments. We wide variety methods used complete five steps localization: (1) define research question, (2) obtain or build time‐synchronizing microphone array, (3) deploy record sounds field, (4) process recordings captured (5) determine location using position estimation algorithms. find eight general purposes ecology behavior systems: assessing individual animals' positions movements, localizing multiple individuals simultaneously study their interactions, determining identities, amplitude directionality, selecting subsets further analysis, calculating species abundance, inferring territory boundaries habitat use, separating from background noise classification. labor‐intensive processing estimating have not yet been automated. In near future, expect increased availability hardware, development automated open‐source software, improvement classification algorithms broaden localization. With three advances, ecologists be better able embrace enabling collection data.

Language: Английский

---

Acoustic indices perform better when applied at ecologically meaningful time and frequency scales

Methods in Ecology and Evolution, Journal Year: 2020, Volume and Issue: 12(3), P. 421 - 431

Published: Oct. 30, 2020

Abstract Acoustic indices are increasingly employed in the analysis of soundscapes to ascertain biodiversity value. However, conflicting results and lack consensus on best practices for their usage has hindered application conservation land‐use management contexts. Here we propose that sensitivity acoustic ecological change fidelity communities negatively impacted by signal masking. Signal masking can occur when responses taxa sensitive effect interest masked less‐sensitive groups, or target sonification is non‐target noise. We argue calculating at ecologically appropriate time frequency bins, effects be reduced efficacy increased. test this a large dataset collected Eastern Amazonia spanning disturbance gradient undisturbed, logged, burned, logged‐and‐burned secondary forests. calculated values two indices: Complexity Index Bioacoustic Index, across entire spectrum (0–22.1 kHz), four narrower subsets spectrum; dawn, day, dusk night. show impact forest classes. Calculating range time–frequency bins substantially increases classification accuracy classes random models. Furthermore, led misleading correlations, including spurious inverse between indicator metrics index compared correlations derived from manual sampling audio data. Consequently, recommend either single narrow bin, predetermined priori understanding soundscape.

Language: Английский

---

Methods for wildlife monitoring in tropical forests: Comparing human observations, camera traps, and passive acoustic sensors

Conservation Science and Practice, Journal Year: 2021, Volume and Issue: 3(12)

Published: Nov. 2, 2021

Abstract Wildlife monitoring is essential for conservation science and data‐driven decision‐making. Tropical forests pose a particularly challenging environment wildlife due to the dense vegetation, diverse cryptic species with relatively low abundances. The most commonly used methods in tropical are observations made by humans (visual or acoustic), camera traps, passive acoustic sensors. These come trade‐offs terms of coverage, accuracy precision population metrics, available technical expertise, costs. Yet, there no reviews that compare characteristics these detail. Here, we comprehensively review advantages limitations three mentioned methods, asking four key questions always important relation monitoring: (1) What target species?; (2) Which metrics desirable attainable?; (3) tools, effort required identification?; (4) financial human resources data collection processing? Given diversity objectives circumstances, do not aim conclusively prescribe particular all situations. Neither claim any one method superior others. Rather, our aims support scientists practitioners understanding options criteria must be considered choosing appropriate method, given their efforts available. We focus on because high priority, although information put forward also relevant other biomes.

Language: Английский

---

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

Scientific Reports, Journal Year: 2021, Volume and Issue: 11(1)

Published: Aug. 3, 2021

Abstract The use of autonomous recordings animal sounds to detect species is a popular conservation tool, constantly improving in fidelity as audio hardware and software evolves. Current classification algorithms utilise sound features extracted from the recording rather than itself, with varying degrees success. Neural networks that learn directly raw waveforms have been implemented human speech recognition but requirements detailed labelled data limited their bioacoustics. Here we test SincNet, an efficient neural network architecture learns waveform using sinc-based filters. Results off-the-shelf implementation SincNet on publicly available bird dataset (NIPS4Bplus) show rapidly converged reaching accuracies over 65% data. Their performance comparable traditional methods after hyperparameter tuning they are more efficient. Learning allows algorithm select automatically those elements best suited for task, bypassing onerous task selecting feature extraction techniques reducing possible biases. We released code datasets encourage others replicate our results apply own datasets; review enhancements hope will become useful bioacoustic tools.

Language: Английский

---

Monitoring the birds and the bees: Environmental DNA metabarcoding of flowers detects plant–animal interactions

Environmental DNA, Journal Year: 2023, Volume and Issue: 5(3), P. 488 - 502

Published: March 8, 2023

Abstract Animal pollinators are vital for the reproduction of ~90% flowering plants. However, many these pollinating species experiencing declines globally, making effective pollinator monitoring methods more important than ever before. Pollinators can leave DNA on flowers they visit, and metabarcoding environmental (eDNA) traces provides an opportunity to detect presence flower visitors. Our study, collecting from seven plant with diverse floral morphologies, eDNA analysis, illustrated value this novel survey tool. using three assays, including one developed in study target common bush birds, recorded animal visiting visual surveys conducted concurrently, bees, other species. We also a visit western pygmy possum; our knowledge is first simultaneously identify interaction insect, mammal, bird flowers. The highest diversity taxa was detected large inflorescence types found Banksia arborea Grevillea georgeana. demonstrates that ease sample collection robustness methodology has profound implications future management biodiversity, allowing us monitor both plants their attendant cohort potential pollinators. This opens avenues rapid efficient comparison biodiversity ecosystem health between different sites may provide insights into surrogate event declines.

Language: Английский

---

Novel community data in ecology-properties and prospects

Trends in Ecology & Evolution, Journal Year: 2023, Volume and Issue: 39(3), P. 280 - 293

Published: Nov. 8, 2023

Language: Английский

---

Effectiveness of acoustic indices as indicators of vertebrate biodiversity

Ecological Indicators, Journal Year: 2023, Volume and Issue: 147, P. 109937 - 109937

Published: Jan. 25, 2023

Effective monitoring tools are key for tracking biodiversity loss and informing management intervention strategies. Passive acoustic promises to provide a cheap effective way monitor across large spatial temporal scales, however, extracting useful information from long-duration audio recordings still proves challenging. Recently, range of indices have been developed, which capture different aspects the soundscape, may estimate traditional measures. Here we investigated relationship between 13 obtained passive estimates various vertebrate taxonomic groupings manual surveys at six sites spanning over 20 degrees latitude along Australian east coast. We found number individual that correlated well with species richness, Shannon's diversity index, total count survey methods. Correlations were typically greater avian than anuran non-avian biodiversity. Acoustic also better richness index. Random forest models incorporating multiple provided more accurate predictions single alone. Out tested, cluster count, mid-frequency cover spectral density contributed greatest predictive ability models. Our results suggest could be tool certain groups. Further work is required understand how site-specific variables can incorporated into improve capabilities taxa besides avians, particularly anurans.

Language: Английский

---

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

Ecological Indicators, Journal Year: 2024, Volume and Issue: 164, P. 112146 - 112146

Published: May 20, 2024

Passive acoustic monitoring has become increasingly popular as a practical and cost-effective way of obtaining highly reliable data in ecological research projects. Increased ease collecting these means that, currently, the main bottleneck ecoacoustic projects is often time required for manual analysis passively collected recordings. In this study we evaluate potential current limitations BirdNET-Analyzer v2.4, most advanced generic deep learning algorithm bird recognition to date, tool assess community composition through automated large-scale data. To end, 3 datasets comprising total 629 environmental soundscapes 194 different sites spread across 19° latitude span Europe. We analyze using both BirdNET listening by local expert birders, then compare results obtained two methods performance at level each single vocalization entire recording sequences (1, 5 or 10 min). Since provides confidence score identification, minimum thresholds can be used filter out identifications with low scores, thus retaining only ones. The volume did not allow us estimate species-specific taxa, so instead evaluated global selected optimized when consistently applied all species. Our analyses reveal that if sufficiently high threshold used. However, inevitable trade-off between precision recall does obtain satisfactory metrics same time. found F1-scores remain moderate (<0.5) studied, extended duration seem currently necessary provide minimally comprehensive picture target community. estimate, however, usage species- context-specific would substantially improve benchmarks study. conclude judicious use AI-based provided represent powerful method assist assessment data, especially duration.

Language: Английский

---