Wildlife Monitoring Research Articles

Improving Object Detection for Time-Lapse Imagery Using Temporal Features in Wildlife Monitoring

Monitoring animal populations is crucial for assessing the health of ecosystems. Traditional methods, which require extensive fieldwork, are increasingly being supplemented by time-lapse camera-trap imagery combined with an automatic analysis of the image data. The latter usually involves some object detector aimed at detecting relevant targets (commonly animals) in each image, followed by some postprocessing to gather activity and population data. In this paper, we show that the performance of an object detector in a single frame of a time-lapse sequence can be improved by including spatio-temporal features from the prior frames. We propose a method that leverages temporal information by integrating two additional spatial feature channels which capture stationary and non-stationary elements of the scene and consequently improve scene understanding and reduce the number of stationary false positives. The proposed technique achieves a significant improvement of 24% in mean average precision (mAP@0.05:0.95) over the baseline (temporal feature-free, single frame) object detector on a large dataset of breeding tropical seabirds. We envisage our method will be widely applicable to other wildlife monitoring applications that use time-lapse imaging.

Overview and Comparison of Deep Neural Networks for Wildlife Recognition Using Infrared Images

There are multiple uses for single-channel images, such as infrared imagery, depth maps, and others. To automatically classify objects in such images, an algorithm suited for single-channel image processing is required. This study explores the application of deep learning techniques for the recognition of wild animals using infrared images. Traditional methods of wildlife monitoring often rely on visible light imaging, which can be hindered by various environmental factors such as darkness, fog, and dense foliage. In contrast, infrared imaging captures the thermal signatures of animals, providing a robust alternative for wildlife detection and identification. We test a Convolutional Neural Network (CNN) model specifically designed to analyze infrared images, leveraging the unique thermal patterns emitted by different animal species. The model is trained and tested on a diverse dataset of infrared images, demonstrating high accuracy in distinguishing between multiple species. In this paper, we also present a comparison of several well-known artificial neural networks on this data. To ensure accurate testing, we introduce a new dataset containing infrared photos of Slovak wildlife, specifically including classes such as bear, deer, boar, and fox. To complement this dataset, the Fashion MNIST dataset was also used. Our results indicate that deep learning approaches significantly enhance the capability of infrared imaging for wildlife monitoring, offering a reliable and efficient tool for conservation efforts and ecological studies.

Transfer Learning for Wildlife Classification: Evaluating YOLOv8 against DenseNet, ResNet, and VGGNet on a Custom Dataset

This study evaluates the performance of various deep learning models, specifically DenseNet, ResNet, VGGNet, and YOLOv8, for wildlife species classification on a custom dataset. The dataset comprises 575 images of 23 endangered species sourced from reputable online repositories. The study utilizes transfer learning to fine-tune pre-trained models on the dataset, focusing on reducing training time and enhancing classification accuracy. The results demonstrate that YOLOv8 outperforms other models, achieving a training accuracy of 97.39% and a validation F1-score of 96.50‘%. These findings suggest that YOLOv8, with its advanced architecture and efficient feature extraction capabilities, holds great promise for automating wildlife monitoring and conservation efforts.

Leveraging local wildlife surveys for robust occupancy trend estimation

Natural resource agencies are frequently tasked with monitoring populations of at-risk species to ensure management activities do not negatively affect the viability of wildlife populations. Typically, these monitoring efforts evaluate trends in a population’s abundance, occupancy, or geographic distribution. Often, surveys provide local information, but results are generally not incorporated into broad-scale monitoring efforts that focus on range-wide population changes due to their variable nature in both spatial extent and effort. We investigated whether aggregating these local (hereafter “variable”) surveys can generate enough statistical power to estimate broad-scale population trends using simulations of declining populations of fishers (Pekaniapennati) over a 10-year time horizon. Our simulations included three population sizes which we refer to as abundant, common, and rare (N0 = 700, 350, and 100 individuals, respectively) with each declining at a rapid and moderate pace (λ = 0.933, and 0.977, respectively). For each population, we simulated variable surveys using an occupancy framework to subsample the population with parameters that mimic combining multiple independent monitoring efforts which vary annually in location, and effort. Regardless of spatial consistency of annual sampling, there was minimal variation in statistical power under both high and low detection probability simulations. However, when sampling effort varied each year, statistical power was lower for most populations and sampling scenarios when compared to consistent sampling effort unless some baseline level of sampling effort was reliably achieved in all years. In many cases, adding low-level consistent baseline sampling to variable surveys resulted in statistical power close to that of consistent sampling efforts. Our results suggest statistical power is driven by annual consistency in the proportion of landscape sampled rather than spatial consistency in sampling locations. This result indicates that current variable surveys could be leveraged and combined to detect population declines for at-risk species at broad-scales if a baseline proportion of landscape is robustly sampled. The level of baseline sampling is highly dependent on population size and magnitudes of population change. In simulations with a common or abundant population experiencing a rapid decline, a baseline survey effort of at least 5% of the landscape in combination with variable surveys resulted in statistical power consistently above the standard threshold of 0.80 for occupancy monitoring. Leveraging existing local efforts to achieve high detection probability and baseline sampling would reduce financial and logistical burdens of broad-scale wildlife monitoring efforts.

Are Domestic Dogs (Canis familiaris) the Family Scapegoats? A Systematic Review of Canine Distemper Virus in African Wildlife, 1978-2021.

Free-living wildlife across Africa is found across expansive rangelands, frequently interacting with pastoral communities, their livestock, and domestic dogs (Canis familiaris). African wildlife populations are threatened by development, poaching and disease. Infectious diseases have caused significant declines, sabotaging conservation efforts. Canine distemper virus (CDV) infections have increased in incidence in wildlife over the past four decades. Sympatric domestic dogs have been presumed to be the reservoirs of the virus for wildlife. A systematic review using Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 guidelines was carried out to investigate CDV infections in wildlife in Africa, to 1) analyze the conservation concerns associated with the disease, 2) identify the dynamics of the virus across different ecoregions, and 3) ascertain the source of the virus for free-living wildlife. The database searches identified 65 articles relevant to the study questions and an additional 43 valuable for wider discussion of the topic area. Canine distemper outbreaks were reported to occur in wildlife in six African countries, with mortality rates between 30% and 94% across all species affected. Eleven wildlife species were identified as susceptible, 64% of which (7/11) are classified as threatened by the International Union for Conservation of Nature. Disease dynamics varied between ecoregions because of differences in land use; virus strains; climate change and concurrent infections; ecotones; domestic and wildlife population densities; wildlife monitoring capacity; and wildlife movement. Nine countries reported outbreaks in domestic dog populations in or near wildlife habitats, but none reported confirmed transmission to wildlife. Of 23 reports investigating the role of domestic dogs as CDV reservoirs, 14 (61%) concluded that domestic dogs acted as either reservoirs or sources of CDV for wildlife, four (17%) did not identify dogs as reservoirs, and five (22%) were not sure of the role played by domestic dogs. This review highlights the importance of integrating active CDV surveillance in wildlife conservation programs.

Inventory and assessment of geosites to promote geotourism in Coto Brus, Costa Rica

This study constitutes a pioneering inventory of geosites and an assessment of Coto Brus, Costa Rica. In collaboration with the Municipality of Coto Brus, we generated baseline knowledge of the existing geosites from this area to support the creation of a touristic route, within the framework of the project One Village, One Product (OVOP), hereinafter OVOP Coto Brus. The slogan of the project is “We are Coto Brus, a paradise Discover us!” The project seeks the economic growth in a region that has been historically dependent on cattle and agriculture. The study reveals the geodiversity of Coto Brus, providing a paleogeographic perspective. On the one hand, it is based on the demonstrated role of geotourism for the socioeconomic development of local economies that adopt it. On the other hand, it contributes to the Sustainable Development Goals (SDG) of the region. Key geosites identified for geotourism include waterfalls such as El Ángel and Santa Fe, the hills of Cerro Pelón and Cerro Quijada del Diablo, and the thermal waters of Aguas Calientes. In addition, the Laguna Zoncho stands out as a lagoon with paleoecological, paleoclimatological, and archaeological importance in ecologically rich sites such as the San Vito Wetland, La Caverna Barranquilla, Cerros Gemelo, and Pittier. They constitute unique opportunities for bird and wildlife observation. Finally, the study recommends that the Municipality of Coto Brus and the stakeholders in the OVOP Coto Brus project implement geosite management plans in the formulation of the touristic route, emphasizing the community's commitment to guarantee sustainable tourism and geoconservation.

Forest 5.0: Internet of Things‐Based Transformation of Forests

ABSTRACTThe phrase “Modernization of forests” highlights the long‐term integration of modern technologies into forest management, aiming to enhance ecosystem observation, data collection, and analysis for exploration and improvement. Technologies like the Internet of Things (IoT), Wireless Sensor Networks (WSN), Artificial Intelligence (AI), and the IoT of Trees are key enablers of this shift. This research explores various technological interventions that could significantly benefit forest ecology, focusing on smart sensing, tracking, and analysis systems for detecting wildfires, illegal logging, and poaching. Simplified structures for tasks like vegetation analysis and forest fire predictions are proposed for future use without altering existing forest management systems. Additionally, the study addresses techniques to improve tribal livelihoods, promote small‐scale forest product marketing, and enhance wildlife monitoring. Key recommendations include establishing reliable wireless communication networks, continuous real‐time monitoring systems, and energy harvesting methods for effective digital forest management.

To crop or not to crop: Comparing whole‐image and cropped classification on a large dataset of camera trap images

AbstractCamera traps facilitate non‐invasive wildlife monitoring, but their widespread adoption has created a data processing bottleneck: a camera trap survey can create millions of images, and the labour required to review those images strains the resources of conservation organisations. AI is a promising approach for accelerating image review, but AI tools for camera trap data are imperfect; in particular, classifying small animals remains difficult, and accuracy falls off outside the ecosystems in which a model was trained. It has been proposed that incorporating an object detector into an image analysis pipeline may help address these challenges, but the benefit of object detection has not been systematically evaluated in the literature. In this work, the authors assess the hypothesis that classifying animals cropped from camera trap images using a species‐agnostic detector yields better accuracy than classifying whole images. We find that incorporating an object detection stage into an image classification pipeline yields a macro‐average F1 improvement of around 25% on a large, long‐tailed dataset; this improvement is reproducible on a large public dataset and a smaller public benchmark dataset. The authors describe a classification architecture that performs well for both whole and detector‐cropped images, and demonstrate that this architecture yields state‐of‐the‐art benchmark accuracy.

Study on noise reduction and structural optimization of ventilated bio-metamaterial plates for acoustic applications

This study focuses on the noise reduction performance and structural optimization of ventilated metamaterial plates designed for bio-acoustic applications, where effective sound attenuation and ventilation are both crucial. Traditional soundproofing materials, which rely on mass and thickness, are inadequate for bio-acoustic environments that require lightweight and compact solutions. In contrast, bio-acoustic metamaterials use resonance effects to attenuate sound while maintaining necessary airflow selectively. This research evaluates multiple metamaterial plate configurations through computational simulations and experimental testing, examining their performance in terms of Sound Transmission Loss (STL), airflow rates, and von Mises stress. The results reveal that Plate Configuration 1 offers the highest STL at 39.14 dB but at the cost of lower airflow efficiency (0.69 m3/s) and increased structural stress (24.83 MPa). Plate Configuration 2 achieves the best airflow efficiency (0.82 m3/s) but with lower noise reduction (STL of 35.42 dB). Plate Configuration 3 provides a balanced performance, with moderate noise attenuation (STL of 37.89 dB), good airflow (0.75 m3/s), and structural stability (von Mises stress of 22.12 MPa). The study concludes that bio-acoustic metamaterials can be effectively optimized for different bio-acoustic applications by carefully tuning their geometry, making them suitable for eco-acoustics, wildlife monitoring, and medical devices where noise control and airflow are critical.

Shadows in the forest: Uncovering unusual colouration records in mammals from the Ecuadorian Tropical Andes.

Variations in colouration patterns have been reported in numerous wildlife species, particularly birds. However, the increased use of camera traps for wildlife monitoring has enabled the detection of elusive species and phenotypic variations that might otherwise go undetected. Here, we compiled records of unusual colouration patterns in terrestrial mammals, documented through camera-trap studies over a 12-year period in the Llanganates-Sangay Connectivity Corridor, in the Tropical Andes of Ecuador. We identified colour variations in seven species of terrestrial mammals, including disorders, such as melanism, white spotting/ piebaldism, xanthocromism and progressive greying. Notably, we reported a high prevalence of melanism in wild populations of the clouded oncilla, along with observations on the species' activity patterns. Approximately half of the recorded clouded oncillas were melanistic. We detected significant differences in activty patterns between melanistic and non-melanistc clouded oncilla, with melanistic morphs showing a peak of activity between 3 a.m. and before dawn. The proportion of melanistic individuals suggests that melanism is widespread throughout the corridor. However, its impact on the species' fitness remains unclear.

SARS-CoV-2 surveillance in captive animals at the belo horizonte zoo, Minas Gerais, Brazil.

The pandemic caused by SARS-CoV-2 has not only affected humans but also raised concerns about its transmission to wild animals, potentially creating natural reservoirs. Understanding these dynamics is critical for preventing future pandemics and developing control strategies. This study aims to investigate the presence of SARS-CoV-2 in wild mammals at the Belo Horizonte Zoo in Brazil, analyzing the virus's evolution and zoonotic potential. The study was conducted at the Belo Horizonte Zoo, Minas Gerais, Brazil, covering a diverse population of mammals. Oropharyngeal, rectal, and nasal swabs were collected from 47 captive animals between November 2021 and March 2023. SARS-CoV-2 presence was determined using RT-PCR, and positive samples were sequenced for phylogenetic analysis. Consensus genomes were classified using Pangolin and NextClade tools, and a maximum likelihood phylogeny was inferred using IQ-Tree. Of the 47 animals tested, nine (19.1%) were positive for SARS-CoV-2. Positive samples included rectal, oropharyngeal, and nasal swabs, with the highest positivity in rectal samples. Three genomes were successfully sequenced, revealing two variants: VOC Alpha in a maned wolf (Chrysocyon brachyurus) and a fallow deer (Dama dama), and VOC Omicron in a western lowland gorilla (Gorilla gorilla gorilla). Phylogenetic analysis indicated potential human-to-animal transmission, with animal genomes clustering close to human samples from the same region. This study highlights the presence of SARS-CoV-2 in various wild mammal species at the Belo Horizonte Zoo, emphasizing the virus's zoonotic potential and the complexity of interspecies transmission. The detection of different variants suggests ongoing viral evolution and adaptation in new hosts. Continuous monitoring and genomic surveillance of SARS-CoV-2 in wildlife are essential for understanding its transmission dynamics and preventing future zoonotic outbreaks. These findings underscore the need for integrated public health strategies that include wildlife monitoring to mitigate the risks posed by emerging infectious diseases.

Automatic identification of the endangered Hawksbill sea turtle behavior using deep learning and cross-species transfer

The accelerometer, an onboard sensor, enables remote monitoring of animal posture and movement, allowing researchers to deduce behaviors. Despite the automated analysis capabilities provided by deep learning, data scarcity remains a challenge in ecology. We explored transfer learning to classify behaviors from acceleration data of critically endangered hawksbill sea turtles (Eretmochelys imbricata). Transfer learning reuses a model trained on one task from a large dataset to solve a related task. We applied this method using a model trained on green turtles (Chelonia mydas) and adapted it to identify hawksbill behaviors like swimming, resting, and feeding. We also compared this to a model trained on human activity data. Results showed an 8% and 4% F1-score improvement with transfer learning from green turtle and human datasets, respectively. Transfer learning allows researchers to adapt existing models to their study species, leveraging deep learning and expanding the use of accelerometers for wildlife monitoring.

Artificial intelligence in environmental monitoring: in-depth analysis

This study provides a comprehensive bibliometric and in-depth analysis of artificial intelligence (AI) and machine learning (ML) applications in environmental monitoring, based on 4762 publications from 1991 to 2024. The research highlights a notable increase in publications and citations since 2010, with China, the United States, and India emerging as leading contributors. Key areas of research include air and water quality monitoring, climate change modeling, biodiversity assessment, and disaster management. The integration of AI with emerging technologies, such as the Internet of Things (IoT) and remote sensing, has significantly expanded real-time environmental monitoring capabilities and data-driven decision-making. In-depth analysis reveals advancements in AI/ML methodologies, including novel algorithms for soil mapping, land-cover classification, flood susceptibility modeling, and remote sensing image analysis. Notable applications include enhanced air quality predictions, water quality assessments, climate impact forecasting, and automated wildlife monitoring using AI-driven image recognition. Challenges such as the “black-box” nature of AI models, the need for high-quality data in resource-constrained regions, and the complexity of real-time disaster management are also addressed. The study highlights ongoing efforts to develop explainable AI (XAI) models, which aim to improve model transparency and trust in critical environmental applications. Future research directions emphasize improving data quality and availability, fostering interdisciplinary collaborations across environmental and computer sciences, and addressing ethical considerations in AI-driven environmental management. These findings underscore the transformative potential of AI and ML technologies for sustainable environmental management, offering valuable insights for researchers and policymakers in addressing global environmental challenges.

Drivers of hunting and photographic tourism income to communal conservancies in Namibia

Hunting and photographic tourism provide ecosystem services that can facilitate conservation. Understanding factors influencing how tourism industries generate income is necessary to ensure sustainable community-based natural resource management. We evaluated effects of large mammal occurrence and landscape attributes on incomes from hunting and photographic tourism earned by communal conservancies in Namibia during 1998–2022. We compiled annual incomes and occurrence of ‘Big 5’ species (elephant [Loxodonta africana], buffalo [Syncerus caffer], black rhino [Diceros bicornis], lion [Panthera leo], and leopard [P. pardus]) using conservancy accounting and wildlife monitoring data. Hunting occurred in 70 of 86 conservancies and generated income almost twice as rapidly as photographic tourism (2.9 and 5.4 years after conservancy establishment, respectively). Hunting income increased with conservancy area and number of Big 5 species present but decreased with conservancy age and increasing mean elevation, topographic diversity, and distances to national parks. Photographic tourism occurred in 39 conservancies and generated 447 % greater median annual income than hunting for conservancies earning >$0. Big 5 species occurrence increased the probability conservancies earned >$0 photographic income but not the amount of photographic income. Photographic income increased with conservancy age and higher annual precipitation but decreased with higher mean elevation. Large mammals are an important driver of income to Namibia’s conservancies and hunting and photographic tourism can provide complementary benefits. We recommend Namibia’s conservancies, particularly those established more recently with smaller area, consider inter-conservancy wildlife co-management and collaboration with tourism industries to improve income potential and develop more sustainable community-based natural resource economies.

A Comparative Analysis of Drone and Boat Monitoring for the Endangered Bolivian River Dolphin

Accurate population monitoring is essential for effective wildlife conservation. This study compares the effectiveness of drone-based vs traditional boat-based methods for assessing an endangered Bolivian river dolphin (Inia geoffrensis boliviensis; BRD) population in Bolivia. Data were collected using high-resolution video recorded with a DJI Mavic 2 zoom drone and standardized methodologies for boat surveys. Two mixed-effects linear models incorporating Poisson and negative binomial error structures were used to compare counts obtained from drone- and boat-based surveys. Results show that drone-based surveys detected 1.15% more individuals on average than boat-based surveys. Drone counts were higher at sites with larger group sizes, leading to congruent estimates. The aerial perspective that drones offer lets researchers overcome potential challenges from boat-based surveys—for example, difficulties related to confirming individual IDs because of limited visibility due to sun glare. Transitioning from boat- to drone-based surveys offers advantages such as reduced disturbance from placement closer to the animals and improved detection rates. Ethical considerations and responsible flight practices are crucial. Standardizing methodologies and prioritizing ethical research factors are key for successful implementation. Drone-based surveys offer a promising approach to enhance wildlife monitoring and conservation practices. This study is the first of this kind for Bolivia, a national natural heritage site, and contributes to the conservation and knowledge of the BRD.

Nasal Carriage of Antimicrobial-Resistant Staphylococci by Fallow Deer (Dama dama) Taken in a Natural Park of Tuscany, Central Italy.

Wild animals are recognized as significant reservoirs for various zoonotic pathogens, including antibiotic-resistant bacteria. This study aimed to investigate the presence of Staphylococcus spp. strains in fallow deer (Dama dama) inhabiting a natural preserve in Central Italy and to examine the phenotypic and genotypic antimicrobial resistance and the presence of some virulence genes among the isolates. During July and December 2022, nasal swabs were collected from 175 fallow deer, which were then analyzed through bacteriological cultures. In total, 176 Staphylococcus spp. strains were isolated and subsequently identified using MALDI-TOF mass spectrometry. S. aureus was the most abundant species with 66 (37.5%) strains, followed by S. hyicus, 34 (19.31%) strains, S. sciuri, 32 (18.18%) strains, S. chromogenes, 27 (15.34%) strains, S. xylosus, 11 (6.25%) strains, S. warneri, 5 (2.84%) strains, and S. devriesei, 1 (0.56%) strain. Antimicrobial susceptibility was assessed for each isolate via the agar disk diffusion method, testing a panel of 13 molecules belonging to 9 antimicrobial classes. The highest resistance rates were detected for penicillin (29.55%), rifampicin (22.73%), and amikacin (20.45%). Notably, intermediate susceptibility was observed for erythromycin (61.93%), enrofloxacin (28.41%), and ceftiofur (21.02%). Conversely, the strains exhibited particularly high susceptibility to amoxicillin/clavulanic acid (99.43%), cefoxitin (97.73%), and vancomycin (96.02%). Based on the results, 32 (18.18%) isolates were classified as multidrug-resistant (MDR). Two strains of S. chromogenes and one strain of S. xylosus, both resistant to penicillin, tested positive for the blaZ gene. No methicillin-resistant strains were found, and none of the isolates harbored genes associated with enterotoxin and toxic shock syndrome toxin production. This study highlights the potential role of wildlife, particularly fallow deer, as reservoirs of antibiotic-resistant Staphylococcus spp. strains. Such findings underscore the importance of monitoring wildlife for antimicrobial resistance, which could have implications for public health and veterinary medicine.

Harnessing Artificial Intelligence for Wildlife Conservation

The rapid decline in global biodiversity demands innovative conservation strategies. This paper examines the use of artificial intelligence (AI) in wildlife conservation, focusing on the Conservation AI platform. Leveraging machine learning and computer vision, Conservation AI detects and classifies animals, humans, and poaching-related objects using visual spectrum and thermal infrared cameras. The platform processes these data with convolutional neural networks (CNNs) and transformer architectures to monitor species, including those that are critically endangered. Real-time detection provides the immediate responses required for time-critical situations (e.g., poaching), while non-real-time analysis supports long-term wildlife monitoring and habitat health assessment. Case studies from Europe, North America, Africa, and Southeast Asia highlight the platform’s success in species identification, biodiversity monitoring, and poaching prevention. The paper also discusses challenges related to data quality, model accuracy, and logistical constraints while outlining future directions involving technological advancements, expansion into new geographical regions, and deeper collaboration with local communities and policymakers. Conservation AI represents a significant step forward in addressing the urgent challenges of wildlife conservation, offering a scalable and adaptable solution that can be implemented globally.

A modeling approach to forecast local demographic trends in metapopulations.

Predicting animal population trajectories into the future has become a central exercise in both applied and fundamental ecology. Because demographic models classically assume population closure, they tend to provide inaccurate predictions when applied locally to interconnected subpopulations that are part of a larger metapopulation. Ideally, one should explicitly model dispersal among subpopulations, but in practice this is prevented by the difficulty of estimating dispersal rates in the wild. To forecast the local demography of connected subpopulations, we developed a new demographic model (hereafter, the two-scale model) that disentangles two processes occurring at different spatial scales. First, at the larger scale, a closed population model describes changes in metapopulation size over time. Second, total metapopulation size is redistributed among subpopulations, using time-varying proportionality parameters. This two-step approach ensures that the long-term growth of every subpopulation is constrained by the overall metapopulation growth rate. It implicitly accounts for the interconnectedness among subpopulations and avoids unrealistic trajectories. Using realistic simulations, we compared the performance of this new model with that of a classical closed population model at predicting subpopulations' trajectories over 30 years. While the classical model predicted future subpopulation sizes with an average bias of 30% and produced predictive errors sometimes >500%, the two-scale model showed very little bias (<3%) and never produced predictive errors >20%. We also applied both models to a real dataset on European shags (Gulosus aristotelis) breeding along the Atlantic coast of France. Again, the classical model predicted highly unrealistic growths, as large as a 200-fold increase over 30 years for some subpopulations. The two-scale model predicted very sensible growths, never larger than a threefold increase over the 30-year time horizon, which is more in accordance with this species' life history. This two-scale model provides an effective solution to forecast the local demography of connected subpopulations in the absence of data on dispersal rates. In this context, it is a better alternative than closed population models and a more parsimonious option than full-dispersal models. Because the only data required are simple counts, this model could be useful to many large-scale wildlife monitoring programs.

Conservation in action: Cost-effective UAVs and real-time detection of the globally threatened swamp deer (Rucervus duvaucelii)

Uncrewed aerial vehicles (UAVs) have proven to be successful tools for ecological monitoring, providing excellent visual resolution and the ability to cover large areas with spatial accuracy. Artificial Intelligence has further improved the capabilities of UAVs vision through object detection. While deep learning has shown significant success in pattern recognition, it still faces challenges in real-world scenarios. In this study, we focused on enhancing the potential of UAVs for wildlife detection and monitoring, specifically focusing on the globally threatened swamp deer (Rucervus duvaucelii). To improve the accuracy of animal recognition with UAVs, we used single-stage detectors YOLO (You Only Look Once) V3, V5, V7, V8, Object detection V3 and DETR (DEtection TRansformer). We trained our model using 48,957 augmented images derived from a dataset of 8210 original true dataset. The result shows the superior performance of Object Detection 3.0 and YOLO V8, achieving a precision score of more than 92 % and a F1 score of more than 85 %, compared to DETR, YOLO V7, V5, and V3. Overall, our study provides an efficient, cost-effective, and accurate detection framework for ecological monitoring that can be non-invasively and least distractively used in various demographic studies of cervids in different regions and habitat types. We have also developed a UAV-based real-time object detection framework that seamlessly integrates with front-end technologies, enabling live detection results regardless of connectivity. This framework operates on a local server, synchronizing with consumer-grade UAVs at a rate of 32 frames per second (fps) with 320 pixel resolution using a frame sampling technique, notably without requiring a dedicated Graphical Processing Unit (GPU). This deliberate choice of not using GPU underscores the commitment to cost-effectiveness and aligns with the research's purpose, prioritizing accessibility and affordability for broader scientific exploration. A least distractive ecological sampling technique was optimized and a maximum of 77 deer were detected and counted in real time within Haiderpur wetland in the Hastinapur Wildlife Sanctuary. This methodology can be replicated and fine-tuned to study other threatened species of conservation priority. This study exemplifies how combining UAVs with deep learning can facilitate species monitoring and population count estimation and be adopted by forest managers to support conservation decisions.

Insights into large carnivore populations in Uganda: A participatory survey of lions, leopards, and hyenas using spatial capture-recapture

Monitoring wildlife populations at scale is fraught with logistical and resource constraints. Despite this, estimating wildlife population state variables and vital rates remains crucial for science, and to assess conservation investment and effort. The veracity and transparency of results also helps prevent politicization of wildlife populations. In Uganda, robust estimates of carnivore population size are rare in the literature. To overcome a near two-decade long data gap we initiated a survey of African lions, leopards, and spotted hyenas with the Uganda Wildlife Authority across six important protected areas. Surveys were conducted within a spatial capture-recapture framework, the industry gold-standard for monitoring carnivore populations. For lions, we used unstructured spatial sampling protocols (in the form of vehicle-based searches), while for leopards, and spotted hyenas we deployed camera traps. Our protocols were designed to obtain unambiguous individual identification photographs. This large-scale effort involved data collection by >100 local conservation stakeholders (including lodge guides, trophy hunters, university students, and government rangers). Locally extinct in three areas, we show lion numbers are precariously low in two of three sites where they still occur (Queen Elizabeth abundance=39.72, Posterior standard deviation (PSD)=7.96; Kidepo Valley abundance=22.23, PSD=11.67). Murchison Falls was identified as Uganda’s lion stronghold with an estimate of ∼240 individuals (PSD=34) and average park-wide densities of 7.43 individuals/100 km2 (PSD=1.05). Leopard densities were highest in Lake Mburo and Murchison Falls. In Murchison they reach some of the highest densities in Africa (14.06 individuals/100 km2, PSD=2.65). Spotted hyena densities were high compared to lions (range=6.15–45.31 individuals/100 km2), except in Lake Mburo where densities were markedly lower (park abundance=23 individuals in 370 km2, PSD=5.26). Our work has critical policy implications, and forms the foundation of the new Strategic Action Plan for Large Carnivore Conservation in Uganda (2023–2033). It also illustrates how cutting-edge, transparent and collaborative science, implemented by multiple wildlife conservation stakeholders can help gauge the conservation status of their wildlife resources.