Biology subjects

Banks-Leite, C.

Publications and source records attributed to Banks-Leite, C..

7 recordsLinked to original sources

bioRxiv2026

Acoustic Monitoring of Tropical Bats: Saccopteryx bilineata's response to habitat and time across a restoration gradient using a 2D-CNN

Global biodiversity loss is a consequence of habitat degradation and fragmentation caused by the expansion of the Anthropocene and the overexploitation of natural resources. Passive acoustic monitoring enables longitudinal assessment of biodiversity but generates massive datasets. Advancements in machine learning have enabled the development of convolutional neural networks capable of automating this process. We developed a lightweight 2D convolutional neural network capable of detecting echolocation calls of the greater sac-winged bat, Saccopteryx bilineata, an insectivorous aerial edge-foraging bat, across a tropical restoration gradient in Para, Brazil. The 2D-CNN model achieved a 97.84% accuracy and 100% precision. Deploying this model across 157,335 field recordings across 29 sampling points resampled over 3 years (2023-2025) yielded 2,098 positive detections of S. bilineata. Generalised linear mixed models revealed that time since reforestation significantly increased the odds of detecting S. bilineata ({beta} = 0.22, p < 0.001), and that detection probability was 2.14 times higher in restoration plots than in forest habitats ({beta} = 0.76, p = 0.0701). Acoustic hardware had a significant covariate influence in both fitted models ({beta} = 0.52, {beta} = 0.72, both p < 0.001). These findings demonstrate the practical applications of deep learning and bioacoustics to quantify habitat recovery and the value of restoration in regenerating functional ecological niches for aerial insectivores during early successional stages.

ecology↗

bioRxiv2026

Towards Automated Acoustic Monitoring of Threatened Birds in Tropical Forest Ecosystem Restoration

Forest restoration has become a widespread response to deforestation and biodiversity loss, particularly across the tropics. Monitoring restoration is essential for tracking progress and requires accurate knowledge of baseline communities, with threatened species often used as recovery indicators. Passive Acoustic Monitoring (PAM) combined with AI classifiers is enabling this at scale. However, such tools remain immature and may underperform for rare, threatened species underrepresented in training data. Here, we demonstrate the potential of PAM assisted by the AI classifier BirdNET+ V3.0, a January 2026 beta release, to support monitoring of threatened tropical birds. By validating 995 detections from 8,940 hours of audio across a reforestation landscape in Para in the Brazilian Amazon, we found that BirdNET+ confirmed the presence of ~43% of the IUCN Red List threatened species possible in the region, five of which were the first confirmed records for the study site. We also found that this method outperformed traditional survey methods. However, BirdNET+ identified species with mixed reliability, and we could also not determine whether restoration has had any impact on them. Our results demonstrate that PAM with BirdNET+ is a viable, practical method for detecting threatened birds in tropical forest restoration landscapes, but that conventional validation approaches are not appropriate for such species, for which extensive validation effort is an essential cost. Under circumstances where this is not possible, we demonstrate validating the single highest-confidence detection per occasion as a practical alternative.

ecology↗

bioRxiv2025

Robotics-assisted acoustic surveys could deliver reliable, landscape-level biodiversity insights

Terrestrial remote sensing approaches, such as acoustic monitoring, deliver finely resolved and reliable biodiversity data. However, the scalability of surveys is often limited by the effort, time and cost needed to deploy, maintain and retrieve sensors. Autonomous unmanned aerial vehicles (UAVs, or drones) are emerging as a promising tool for fully autonomous data collection, but there is considerable scope for their further use in ecology. In this study, we explored whether a novel approach to UAV-based acoustic monitoring could detect biodiversity patterns across a varied tropical landscape in Costa Rica. We simulated surveys of UAVs employing intermittent locomotion-based sampling strategies on an existing dataset of 26,411 hours of audio recorded from 341 static sites, with automated detections of 19 bird species (n = 1,819) and spider monkey (n = 2,977) vocalisations. We varied the number of samplers deployed in a single survey (sampling intensity) and whether the samplers move between sites randomly, in a pre-determined route to minimise travel time, or by adaptively responding to real-time detections (sampling strategy) and measured the impact on downstream ecological analyses. We found that observed avian species richness and spider monkey occupancy was not impacted by sampling strategy, but that sampling intensity had a strong influence on downstream metrics. Whilst our simulated UAV surveys were effective in capturing broad biodiversity trends, such as spider monkey occupancy and avian habitat associations, they were less suited for exhaustive species inventories, with rare species often missed at low sampling intensities. As autonomous UAV systems and acoustic AI analyses become more reliable and accessible, our study shows that combining these technologies could deliver valuable biodiversity data at scale.

ecology↗

bioRxiv2023

Automatic vocalisation detection delivers reliable, multi-faceted, and global avian biodiversity monitoring

Tracking biodiversity and its dynamics at scale is essential if we are to solve global environmental challenges. Detecting animal vocalisations in passively recorded audio data offers a highly automatable, inexpensive, and taxonomically broad way to monitor biodiversity. However, uptake is slow due to the expertise and labour required to label new data and fine-tune algorithms for each deployment. In this study, we applied an off-the-shelf bird vocalisation detection model, BirdNET, to 152,376 hours of audio comprising of datasets from Norway, Taiwan, Costa Rica, and Brazil. We manually listened to a subset of detections for each species in each dataset and found precisions of over 80% for 89 of the 139 species (100% for 57 species). Whilst some species were reliably detected across multiple datasets, the performance of others was dataset specific. By filtering out unreliable detections, we could extract species and community level insight on diel (Brazil) and seasonal (Taiwan) temporal scales, as well as landscape (Costa Rica) and national (Norway) spatial scales. Our findings demonstrate that, with a relatively fast validation step, a single vocalisation detection model can deliver multi-faceted community and species level insight across highly diverse datasets; unlocking the scale at which acoustic monitoring can deliver immediate applied impact.

ecology↗

bioRxiv2023

Constraints on avian seed dispersal reduce potential for resilience in degraded tropical forests

1. Seed dispersal - one of the many services supplied by biodiversity - is a critical process underpinning the resilience of tropical forests. Forest loss or degradation typically leads to defaunation, altering seed transfer dynamics and impairing the ability of forested habitats to regenerate or recover from perturbation. However, the extent of defaunation, and its likely impacts on the seed dispersers needed to restore highly degraded or clear-felled areas, remains poorly understood, particularly in human-modified tropical forest landscapes. 2. To quantify defaunation of seed-dispersing birds, we used field survey data from more than 400 transects in three regions of Brazil, first comparing the recorded assemblages with those predicted by geographic range maps, and then assessing frugivore habitat associations across gradients of land cover modification at local scales. 3. We found that current bird assemblages have lower functional trait diversity than predicted by species range maps in Amazonia (4-6%), with a greater reduction (28%) for the Atlantic Forest region, which has been more heavily deforested for a longer period. These reductions are probably caused by local extinctions of forest-dependent bird species following land-use change. 4. Direct measures of seed dispersal are difficult to obtain, so we instead focused on the potential for seed transfer inferred from shared species occurrence between land cover types. Of 83 predominantly frugivorous bird species recorded in relatively intact forests, we show that 10% were absent from degraded forest, and 57% absent from the surrounding matrix of agricultural land covers, including many of the large-beaked species. Of 112 frugivorous species using degraded forest, 47% were absent from matrix habitats. 5. Our findings suggest that degraded forest can supply seed dispersal services to adjacent cleared lands, and that direct transfer of seeds from intact forest to cleared areas may be limited, particularly for large-seeded trees. We conclude that resilience of tropical forest landscapes is best achieved by protecting a mosaic of forest types, including sufficient core areas of intact forest surrounded by buffer zones of degraded forest.

ecology↗

bioRxiv2023

Mangrove propagules are limited in their capacity to disperse across long distances

Mangroves are subject to rapid and large-scale habitat changes which threaten their unique genetic diversity and provision of critically important ecosystem services. Habitat fragmentation reduces connectivity which can impair dispersal and lead to genetic isolation. However, it is unclear whether fragmentation could impact mangrove genetic isolation, as mangrove propagules can disperse long distances. Here, we conducted a meta-analysis of studies reporting a correlation between geographic distance and genetic distance in mangrove plants. From the 22 studies that met the inclusion criteria, we found a significant isolation-by-distance effect; geographic distance was significantly associated with Neis genetic distance and FST. Our results show that mangrove propagules may be limited in their capacity to disperse across long distances, which highlights the importance of maintaining close proximity between habitat patches and reducing habitat fragmentation.

ecology↗

bioRxiv2020

Disentangling the effects of sampling scale and size on the shape of species abundance distributions

Many authors have tried to explain the shape of the species abundance distribution (SAD). Some of them have suggested that sampling scale is an important factor shaping SADs. These suggestions, however, did not consider the indirect and well-known effect of sample size, which increases as samples are combined to generate SADs at larger scales. Here, we separate the effects of sample size and sampling scale on the shape of the SAD for three groups of organisms (trees, beetles and birds) sampled in the Brazilian Atlantic Forest. We compared the observed SADs at different sampling scales with simulated SADs having the same richness, relative abundances but comparable sample sizes, to show that the main effect shaping SADs is sample size and not sampling scale. The effect of scale was minor and deviations between observed and simulated SADs were present only for beetles. For trees, the match between observed and simulated SADs was improved at all scales when we accounted for conspecific aggregation, which was even more important than the sampling scale effect. We build on these results to propose a conceptual framework where observed SADs are shaped by three main factors, in decreasing order of importance: sample size, conspecific aggregation and beta diversity. Therefore, studies comparing SADs across sites or scales should use sampling and/or statistical approaches capable of disentangling these three effects on the shape of SADs.

ecology↗