Remote Sensing in Ecology and Conservation

Waterbirds serve as important indicators of both aquatic and terrestrial ecosystem health, making effective monitoring essential for tracking population health and identifying potential causes of decline. Drones have provided opportunities to overcome historic waterbird monitoring challenges, but the expertise and time required for manual image analysis creates a major bottleneck. Recent advances in deep learning-based object detection have enabled rapid, automatic detection of features in complex ecological imagery, though applications have largely been limited to single-species colonies, and practitioners lack quantitative comparisons of annotation time and accuracy across different levels of automation. We systematically compared four waterbird monitoring approaches using identical survey areas from Chester Island, a mixed-species colony in Matagorda Bay, Texas, in 2025: (1) traditional ground-based counts, (2) manual drone imagery-based counts, (3) computer-assisted counts using pre-annotations from an object detector with manual human verification (Human+ML), and (4) fully automated counts using object detector annotations (ML-only). We trained a YOLOv10 object detection model on manually annotated imagery of Chester Island in 2021 and applied it to the 2025 imagery. Manual drone annotation detected 6,530 birds in 40.5 hr and served as the primary reference standard. Human+ML detected 5,826 birds (89% of manual) in 7.7 hr, an 81% reduction in annotation time. ML-only detected 5,679 birds (87% of manual) in approximately 46 min, a 98% reduction. Ground counts recorded 5,868 birds (90% of manual). Detection generalized well across species while classification depended heavily on training data and morphological distinctiveness. The Human+ML workflow emerged as a practical middle ground, providing practitioners with empirical data to evaluate partial versus full automation strategies based on monitoring objectives.

Wildlife re-identification (re-ID) is a widely used and powerful tool with diverse applications in animal ecology and conservation. Current automated methods typically operate on single images of a single body part of the animal. However, a single encounter may contain multiple images capturing different body regions, each providing complementary individual-specific information. In contrast to automated approaches, researchers often manually select the most suitable images and regions for identification based on factors like visibility, occlusion and image quality. This creates a mismatch between automated methods and field practice, limiting the practical adoption of current automated re-ID pipelines. Here, we address this by introducing an encounter-based, multi-body-part re-ID framework, using sea turtles as a model taxon. Our framework combines three elements: (1) An orientation-aware deep learning model, TurtleDetector, that in addition to the full bodies, it also automatically segments key body regions, i.e. heads, front and hind flippers, from images within an encounter; (2) a hybrid body-part-specific retrieval method, that sequentially combines a fast global-feature model (MiewID or DINOv3) with a more accurate but costlier local-feature model (ALIKED with LightGlue); and (3) a merged identity-prediction strategy that selects the highest calibrated similarity score across all available body parts and images of an encounter. We evaluate the framework on three long-term re-ID datasets spanning three species, loggerheads, greens, and hawksbill turtles, under an evaluation protocol that mirrors real-world, time-aware re-ID workflows. Across datasets, combining multiple body regions consistently improved identification performance over the best-performing single body region, resulting to an increase of 4-6% in top-1 accuracy. Interestingly, body regions traditionally underused in sea turtle re-ID, such as the hind flippers and carapaces, provided complementary identifying information that improved encounter-level re-ID when integrated through the hybrid retrieval method. Our findings demonstrate that automated wildlife re-ID can benefit from moving beyond single-image, single-body-part identification towards encounter-level integration of all available visual evidence. Our work further suggests that, where feasible, field photo-acquisition protocols should aim to capture multiple informative views of an individual during each encounter. Importantly, many species and taxa, including elephants, primates, cetaceans, and other large vertebrates, possess such individual-specific features across multiple body regions, highlighting the broad potential applicability of our framework.

Camera trap networks such as Snapshot Safari have generated millions of labelled wildlife images across Africa, enabling the training of deep learning models for automated species classification. However, deploying models trained in one African region to another remains poorly understood. To the best of our knowledge, this study presents the first systematic evaluation of geographic domain shift within the African continent for wildlife camera trap species classification, using the Machine Learning sub-field of Artificial Intelligence. We use three model architectures, each interacting with Snapshot Serengeti in a different way: BEiTV2is fine-tuned on Serengeti images as a supervised baseline; DINOv2 with FAISS uses Serengeti images as a retrieval index without any weight updates; and BioCLIP is a true zero-shot foundation model that receives no Serengeti training data at all. All three are then evaluated on two Southern African test sets, Snapshot Kgalagadi and Snapshot Kruger, as well as on locally collected wildlife photographs from Botswana. We conduct eight experiments covering in-domain baselines, cross-dataset transfer, data scaling, MegaDetector preprocessing, grayscale vs. colour image conditions, and per-species transfer analysis. This work provides the first empirical characterisation of intra-African domain shift across both supervised and zero-shot architectures, and offers practical guidance for conservation AI practitioners who need to deploy models across the diverse ecosystems of Southern Africa without collecting new labelled data.

High-resolution monitoring of forest structure and productivity is essential for effective natural resource management. However, monitoring approaches such as field-based forest inventories or extensive lidar campaigns are costly, time-intensive, and spatially limited. Therefore, inexpensive and accessible methods are needed. SatCHM (Satellite Canopy Height Model) was developed to be an accessible and open-source tool for researchers, allowing for site-specific and temporally flexible predictions of canopy height with limited computational resources. SatCHM requires four inputs: panchromatic satellite imagery, solar and sensor angle metadata of satellite imagery, digital elevation models (DEMs), and lidar-produced CHMs for an area of interest. After SatCHM pre-processes inputs, data is loaded into a collection of convolutional neural networks (CNNs) for image-to-image regression. This ensemble cooperates to yield high-resolution predictions (up to 0.5-meter) of three-dimensional tree structure with discernible tree crowns across a broader defined area of interest. After calculating the mean absolute error for each prediction output, the median of these mean absolute errors was 6.06 meters.

Coral reef fish assemblages are widely used as indicators of ecosystem condition, yet manual annotation of underwater video remains a major bottleneck for scalable biodiversity monitoring. Despite rapid progress in automated detection, ecologically realistic and publicly available datasets remain scarce, particularly for the Western Indian Ocean. Here, we present WIO-ReefFish, a reef fish detection dataset derived from diver-operated line-intercept transects and designed for ecological monitoring under natural survey conditions. WIO-ReefFish comprises 1,000 ultra-high-definition images (3840 $\times$ 2160 pixels) and 6,768 exhaustive bounding-box annotations spanning 24 taxonomic categories, thereby preserving full-frame assemblage structure in complex reef scenes. We also establish a standardized benchmark across nine object detection models under two complementary protocols: class-aware detection and class-agnostic fish localization. Detection performance was consistently higher under the class-agnostic protocol. The best-performing model (RT-DETR) improved from 0.48 mAP50 in the class-aware setting to 0.70 mAP50 when taxonomic constraints were removed, indicating that taxonomic discrimination remains substantially more challenging than fish localisation in reef imagery. Spatially independent evaluation revealed a pronounced generalisation gap, particularly for taxonomic detection, whereas class-agnostic fish localisation remained substantially more robust across transects and countries. Together, these results establish WIO-ReefFish as a realistic benchmark for automated reef fish detection and provide a foundation for more robust computer-vision tools in coral reef biodiversity monitoring. The WIO-ReefFish dataset and associated benchmarking resources are publicly available.

Automated monitoring of insect pollinators in natural environments with insect camera traps and trained deep learning algorithms provides novel data for insect ecological studies. However, efficient and accurate image recognition analysis of the recorded images or videos is challenging, particularly for images containing small insects against complex backgrounds with diverse vegetation communities. Even when insects can be detected in images, identifying their taxonomy remains difficult, particularly in footage with low image resolution, light conditions, and distances from the plants, and in cases where insects appear blurry or only partially visible. In this work, we present InsectDCT, an AI-based pipeline for automated detection, hierarchical classification, and tracking of insects in footage of natural vegetation tested in different environments. The InsectDCT pipeline consists of three levels: insect Detection and localization, hierarchical taxonomic Classification, and spatio-temporal Tracking. In the first stage, insects are detected in time-lapse images or video recordings using the You Only Look Once (YOLO11) object detection architecture. Detection performance is improved using motion-enhanced images, which improve robustness in cluttered and 3 dimensional environments. The detector is trained on an extensive dataset that contains more than 60,000 images collected using camera traps deployed across a wide range of plant families and floral habitats. In the second stage, detected insects are classified using a hierarchical taxonomy-aware classification framework that covers 80 taxonomic groups. Classification is performed at multiple taxonomic levels, including order, family, and genus/species, allowing coarse and fine-grained ecological analyzes while accounting for varying levels of visual ambiguity. In the third stage, a multi-object tracking module is applied to high temporal-resolution image sequences and video data to associate detections of the same individual across time. InsectDCT code and all datasets are made publicly available. Author summaryInsects are declining worldwide, creating an urgent need for efficient methods to monitor their abundance, activity, and diversity. Traditional insect surveys often require extensive fieldwork and expert taxonomic identification, which limits the scale and frequency of monitoring. In this study, we developed InsectDCT, an artificial intelligence-based pipeline that automatically detects, classifies, and tracks insects in camera-trap recordings collected from natural and semi-natural environments. Our approach combines deep-learning methods for object detection, hierarchical taxonomic classification, and tracking of individual insect observations through time. Unlike many existing systems that are trained for a single habitat or plant species, we designed our framework using images collected across a wide range of flowering plants, camera systems, and insect groups. This makes the system more transferable to new ecological settings. The classifier can identify insects at multiple taxonomic levels and can return higher-level classifications when species-level identification is uncertain. We demonstrate that the pipeline can process large image datasets efficiently, including on low-power edge-computing devices such as Raspberry Pi systems. By providing both the software and the underlying datasets, we aim to support scalable, non-invasive insect monitoring and facilitate future ecological and conservation research.

Individual southern right whales (Eubalaena australis) can be identified from the callosity pattern on the head, a stable pattern of roughened keratinized patches. Manual comparison of drone images with large catalogs, however, is time-consuming. For southern right whales, automated photo-identification approaches typically require substantial training data or retraining when new individuals are added. Here we evaluate two classical image-similarity methods for individual identification from standardized dorsal head images: histograms of oriented gradients (HOG), and symmetric log-chamfer distance. We tested 375 query images from 198 known whales with recurrent sightings against 411 reference images, one for each individual whale. HOG ranked the correct whale first in 361 of 375 cases (96.3%), whereas log-chamfer ranked the correct whale first in 355 cases (94.7%). All incorrect rank-1 matches could be identified by a high risk score computed from query-reference distance distributions. A combined rule selecting the first-ranked candidate from the method with the lower risk score increased rank-1 accuracy to 368 cases (98.1%). These results show that classical registered image matching provides a practical tool for southern right whale photo-identification.

Community science platforms like iNaturalist generate unprecedented volumes of biodiversity data, but their scientific utility depends critically on accurate species identification--a persistent challenge when contributors often lack taxonomic expertise. We developed "LizardLens", a two-stage machine learning pipeline that decouples object detection from species classification to enable fine-grained identification of morphologically similar organisms in visually complex field photographs. Using 10,000 verified iNaturalist images of five Anolis lizard species in Florida, we trained specialized YOLO-based detection and Swin Transformer classification models and compared performance against state-of-the-art single-stage architectures. Our two-stage pipeline achieved 83.0% Top-1 accuracy and a macro-averaged F1-score of 89.0%, indicating strong precision-recall performance across species and outperforming single-stage YOLOv8 and YOLOv12 models across all evaluation metrics for all species, with relative improvements ranging from 10.5% to 13.2%. Gradient-weighted Class Activation Mapping (Grad-CAM) indicated that the models predictions were consistently associated with regions corresponding to diagnostic morphological (e.g., head shape, feet, and limb lengths) and pattern features (e.g., ocular rings and body patterning), providing evidence that LizardLens leverages biologically relevant visual cues consistent with those used by expert taxonomists. Error analysis identified partial occlusion and multiple proximate individuals as primary sources of missed detections, while spurious detections of lizard-like environmental features (e.g., sticks, bark) represented the dominant false positive error mode. We deployed LizardLens as an accessible web application featuring interactive bounding box correction, ranked species predictions with confidence scores, directly supporting the "Lizards on the Loose" middle school community science initiative. By combining technical advances in fine-grained visual classification with user-centered design, LizardLens demonstrates how machine learning can simultaneously enhance data quality for biodiversity monitoring and provide authentic scientific experiences for student participants. Our approach is generalizable to other small-bodied organisms in complex habitats and provides a framework for translating computer vision advances into practical tools for community science and conservation.

The US NEXRAD radar network has monitored the aerosphere over the US and its territories continuously since the 1990s and archived nearly 300 million radar volume scans. These data contain a wealth of information about the movements of birds, bats, and insects. Historically, this biological information was difficult to access due to the amount of data and challenges in analyzing it. In the last 15 years, fueled by computational and methodological advances, large-scale aeroecology research has blossomed. However, comprehensive analyses of the NEXRAD archive remain very costly. We collected measurements of biological activity from every volume scan in the NEXRAD archive--nearly 300 million data files total--to assemble a dataset of aerial biomass over the US from 1995 to 2025. The core data are vertical profiles, which summarize biological activity at different heights above the radar station for each volume scan. We also provide time series data products that aggregate vertical profiles to point measurements at radar stations across time. These data products can support a range of aeroecology analyses at significantly reduced effort.

Fine-grained localisation of plumage regions is a prerequisite for computational analyses of avian colouration, patterning and visual traits in ecological and evolutionary research. Progress is limited by the scarcity of image resources with annotations aligned to biologically meaningful anatomical units: existing avian benchmarks provide either landmark points or coarse part categories that do not capture ornithologically defined plumage regions. We present a curated dataset of 4,705 bird images annotated for nine plumage regions: head, throat, breast, belly, vent, back, coverts, remiges and tail. Spanning 39 avian orders and 222 families, the dataset provides a taxonomically broad resource for fine-grained avian image analysis. The dataset was built through an iterative model-assisted annotation workflow, in which model predictions were reviewed and corrected rather than drawn from scratch, improving the efficiency of region-level annotation. We benchmark classical segmentation architectures, SAM-based models and self-supervised foundation-model encoders on this task. A frozen DINOv3 encoder with a lightweight decoder achieved the highest performance on the held-out test set, reaching 84.01% mean Intersection over Union while requiring substantially less memory than end-to-end fine-tuning. The model generalised to external avian benchmarks, including the bird subset of PartImageNet and CUB-200-2011, and achieved competitive performance on the full PartImageNet part-segmentation benchmark, which includes diverse animal taxa. We provide a modular detect-track-segment pipeline as a proof-of-concept extension to video data. Together, these results show that anatomically grounded avian annotations can serve both as a resource for plumage phenotyping and as a benchmark for efficient, transferable biological part segmentation. Author SummaryBirds vary enormously in colour and pattern, but studying this variation at large scales requires more than identifying the bird in a photograph. Researchers often need to know where each colour or pattern occurs on the body, such as on the head, throat, breast, wing or tail. We created PlumageParts to make this kind of region-level analysis easier. The dataset contains 4,705 bird images annotated into nine biologically meaningful plumage regions, covering a wide range of bird families and orders. To build the dataset efficiently, we used a model-assisted workflow in which computer-generated masks were checked and corrected by researchers rather than drawn entirely by hand. We then tested several image-segmentation approaches and found that a frozen self-supervised vision model, combined with a lightweight decoder, provided accurate plumage-region predictions while requiring relatively modest computing resources. The same approach also performed well on a broader animal part-segmentation benchmark, suggesting that it may be useful beyond birds when suitable annotations are available. By releasing the annotations, code and trained model, we aim to support future studies of bird plumage and biologically meaningful image segmentation.

Mosquito-borne diseases represent an escalating global health threat, driven by climate change, urbanization, and the spread of invasive vectors into new territories. Effective surveillance is constrained by a critical taxonomic impediment: the rate of specimen collection far outpaces the capacity of expert entomologists to process and identify trap catches. To address this bottleneck we developed MosAICo, an integrated AI-powered ecosystem for automated mosquito species identification designed for real-world, national-scale entomological surveillance. The system combines a standardized benchtop imaging device with MosAICo-Net, a deep learning pipeline enabling efficient and principled open-set recognition and uncertainty quantification. Trained and evaluated on 12, 499 specimens spanning 15 species collected across Italy, the model identifies seven priority vector species while explicitly rejecting out-of-distribution specimens. On a geographically stratified held-out test set, MosAICo-Net achieved over 90% accuracy on target species, and an AUROC of 0.96 for out-of-distribution detection. Field validation across 20 Italian surveillance sites confirmed these results: 94% micro accuracy on 1, 470 field-collected target specimens and strong agreement with expert manual counts ([Formula] = 0.66). To assess cross-geographic generalizability, the system was further evaluated on 118 Aedes albopictus specimens collected at the fringe of the species invasion front in Ghana: a 97.4% accuracy with only a single specimen escalated to expert review, suggests that MosAICo is well-suited for deployment in distant and epidemiologically critical regions. The system processes up to 82 specimens per image, matching expert throughput at constant speed regardless of taxonomic complexity. By embedding uncertainty-aware AI within a standardized hardware-software pipeline, MosAICo acts as a scalable force multiplier for public health entomology, freeing expert attention for rare, invasive, or ambiguous specimens that require human validation.

Coastal forests are increasingly threatened by saturated soil and elevated salinity levels resulting from sea level rise, saltwater intrusion, and storm surges. In response to rising salinization and flooding, healthy coastal forests that rely on freshwater (both wetland forests and low-elevation upland forests) are transitioning into landscapes dominated by dead or dying trees, known as ghost forests. Situated among salt-tolerant shrubs and grasses, ghost forests eventually become marshes or open water. Here, our main objective was to quantify the dynamics and pathways of these forest landscape conversions, as well as the factors contributing to the changes, which is vital for understanding the progression of coastal ecosystem degradation and forecasting future changes. We focused first on identifying the best method to track forest landscape change by exploring the role of multiple remote sensing indices (i.e., multispectral, bi-seasonal, topographical, and phenological metrics) in enhancing the performance of deep learning models (convolutional neural networks, CNNs) for land cover classification in the coastal plain of North Carolina using surface reflectance of Landsat 8 and Sentinel-2 images. Then, we used the best available data (Landsat 8) to understand long-term change and identify patterns of land cover change from 1985 to 2021. Our study reveals that incorporating phenology and topographical indices enhances the separability of the ghost forests class from all other vegetation classes. In our assessment, the higher-resolution Sentinel-2 data (F1 Score = 96.3) outperformed Landsat images (F1 score = 93.4) for the 2021 co-available year. However, Landsat remains an important tool used due to its long-term data record. Therefore, we used Landsat to determine that 21% of forests were lost between 1985 and 2021, and that the rate of loss is increasing. Between 2010 and 2021, 23,876 ha of forest were converted to marsh, ghost forest, and shrub, which is 1.5 times higher than the 16,968 ha lost between 1985 and 2010. These conversions from forest to ghost forest and marshes were driven primarily by proximity to the channel, salinity, and the increasing rate of relative sea level rise (RSLR), which are the key environmental drivers of observed changes. By quantifying these changes, we highlight regions most vulnerable to environmental stressors, providing a basis for targeted conservation strategies.

Banana diseases impose severe production losses in tropical smallholder farming systems, yet accurate in-field visual diagnosis remains difficult: symptom expression varies across cultivars and growth stages, and several diseases produce morphologically overlapping foliar signs. We developed a probabilistic image-recognition framework for detecting five economically important banana diseases -- Xanthomonas Wilt, Banana Bunchy Top Disease, Fusarium Wilt (Panama disease), Yellow Sigatoka, and Black Sigatoka -- from in-field photographs, without any disease-specific fine-tuning of the vision backbone. The approach extracts frozen 1,152-dimensional embeddings from the DINOv3 vision foundation model and couples them with a conditional normalizing flow, trained on four publicly available datasets spanning diseased banana plants, healthy tissue, non-banana vegetation, and general natural imagery. On an independent test set the model achieved F1 scores exceeding 0.98, average precision values of 0.968-0.999, and AUROC values of 0.997-1.000 across all five diseases evaluated as binary detection problems. Multi-class accuracy was near-perfect, with limited confusion between Yellow Sigatoka and Black Sigatoka -- a biologically plausible ambiguity attributable to overlapping early-infection foliar symptoms. Because the normalizing flow estimates explicit conditional probability densities rather than decision boundaries, two complementary log-likelihood ratios can be derived: a disease ratio comparing each disease class against healthy banana, and a plant ratio comparing banana against non-banana imagery. Together these define an interpretable two-dimensional diagnostic space that simultaneously quantifies evidence for disease presence and image relevance, cleanly separating diseased plants, healthy plants, and out-of-distribution images while flagging uncertain predictions for confirmatory testing. Inference on frozen embeddings is lightweight and compatible with smartphone deployment, providing a scalable, uncertainty-aware diagnostic tool for smallholder farming systems and disease surveillance programmes.

Contextualizing the movements of animals into their three-dimensional (3D) habitat contexts is still a major challenge for fields relying on animal tracking methods. This promises to change as Structure-from-Motion photogrammetry tools and techniques revolutionize image processing into the truly 3D spatial realm, by enabling reconstructions of habitat models from overlapping photographs. Combined with tracking data, these techniques would help elucidate drivers behind animal movements that have been previously masked by two-dimensional approaches. Unfortunately, tracking methods are often still impractical for use with understudied or non-model animals, especially those living underwater. In this paper, we describe a method for tracking the translational movements of animals into a photogrammetric habitat model. Our approach spans three general parts: (1) filming the navigation paths of wild animals by following individuals with small cameras (GoPros) on extendable sticks whilst SCUBA diving, (2) reconstructing a 3D spatial habitat model from separate footage, and (3) manually plotting the 3D trajectories of animals into the habitat model. We used one popular commercial software for photogrammetric reconstruction, trajectory plotting, and measurement of trajectories, after which the plots can be exported in a variety of formats for further analyses. Straightforward and flexible methodologies such as this stand to encourage more fieldwork concerning animals that live in structurally complex habitats, or animals that are underrepresented in movement or navigation research. We expect that this approach can be adapted to study many aquatic or terrestrial animals in different habitats, and at various scales.

The tidal wetlands of Louisiana comprise about 25% of those found throughout the conterminous United States yet estimates of wetland loss rates in the region between 1932 and 2016 have exceeded 60 km2 yr-1. To mitigate further degradation and wetland loss in the region, a globally unprecedented $50B, 50-year plan for coastal Louisiana is driving restoration efforts, and demand exists from multiple stakeholders for regularly updated, regional-scale, accurate land cover information. We used machine learning (random forests; RF) and cloud computing to develop a new Landsat-based, marsh vegetation community geospatial dataset. The dataset depicts wetland vegetation community types defined in a previous study at annual (1985-2025) time steps at 30-m resolution. An RF algorithm was used to integrate training samples with feature variables derived from Landsat imagery, and the resulting geospatial data product achieved an overall correct classification rate of 78%. The approach for development of the land cover dataset presented here has potential for application in other coastal wetland habitats throughout the world.

1. Passive acoustic monitoring (PAM) is increasingly used for ecological research, biodiversity monitoring, assessment, and reporting. Automated species classifiers make it feasible to process large audio datasets but generate numerous detections that often need validation before use in downstream analyses or formal outputs. 2. Method development in PAM has focused on classifier building and downstream models that account for imperfect detection, yet the practical step between these - post-classification validation - remains weakly supported and is often implemented through ad hoc workflows. This increases manual handling, creates scope for transcription or consolidation errors, limits transparency and makes it difficult to document what was reviewed. 3. We introduce PAMalytics, an open-source, no-code, local browser-based application to support post-classification validation as a standardised workflow stage. PAMalytics ingests detections from any classifier, allows users to define how detections are sampled for review, and presents selected detections alongside their spectrograms with audio playback in one unified interface. Sampling strategy and review decisions are tracked alongside reviewer identity improving traceability and reproducibility across the validation workflow. 4. Case studies with Conservation International Cambodia and Imperial College London demonstrate PAMalytics in two validation settings. In Cambodia, gibbon predictions from a large, uneven dataset were sampled within sites, with likely classifier errors prioritised for validation. At Imperial, Amazon bird detections were sampled across each species classifier-confidence range before biodiversity metrics were derived. In both cases, PAMalytics reduced manual handling and validation time. By turning an ad hoc step into an accessible, structured workflow for conservation practitioners, PAMalytics fills a practical gap in the PAM bioacoustics pipeline and strengthens the link between automated detections and evidence used in biodiversity monitoring and reporting.

Models for estimating animal density from camera traps require four parameters informing detection: movement speed, daily activity level, staying time (duration animals remain within the detection zone), and effective detection distance. These parameters traditionally come from labor-intensive manual measurements and auxiliary telemetry. Recent advances in computer vision can provide the positions of animals in camera trap images, which have been used for distance sampling. We extend this approach to extract all four parameters from imagery, providing the first AI-derived estimates of movement speed and staying time from automated coordinate tracking. We also introduce a new joint multi-species hierarchical distance function that estimates deployment-specific effective detection distances while borrowing strength across species through partial pooling. Our pipeline integrates MegaDetector for animal detection, the Segment Anything Model for segmentation, and Dense Prediction Transformers for monocular depth estimation. From frame-level coordinates, we reconstruct movement trajectories across burst sequences to estimate speed with size-biased distribution corrections, calculate staying time through bounding box interpolation, and estimate activity levels from detection timestamps. The joint hierarchical distance function decomposes the detection scale parameter into a shared deployment-level effect and species-specific offsets, so species effects represent deviations from the multi-species average, allowing data-rich species to inform detection conditions where rare species have few observations. AI-derived scene depth enters the model as a covariate on detection range, providing a vegetation openness metric from the same pipeline. To address position errors from depth estimation, we apply data quality filters. We processed 122,574 frames from 181 deployments across montane forests in Washington and Montana, generating parameter estimates for 12 species without manual annotation. Automated speed estimates produced day ranges 2.7 to 4.3 times GPS telemetry-derived daily distances, reflecting differences between encounter velocity within detection zones and landscape-scale displacement. Deployment-level variation in detectability exceeded species-level differences 3:1, with scene depth strongly predicting detection range; mean effective detection distances ranged from 4.1 to 7.6 m. Applied to a Random Encounter Model, these parameters yielded a white-tailed deer density estimate of 21.4 animals/km{superscript 2} and the Random Encounter Staying Time model yielded 11.6animals/km{superscript 2} in Montana. This pipeline enables scalable density estimation across large camera trap networks.

O_LISolitary bee species that use artificial trap nests are important for agricultural crop production and as indicators of habitat quality. Quantifying cavity-nesting solitary bee foraging and nesting behavior is essential for real-time analysis of population numbers and pollination activity, as well as understanding how environmental conditions shape reproductive success and population dynamics. However, manual observation is labor-intensive, prone to observer bias, and unable to deliver continuous data. Existing automated systems either require individual bee marking or detect presence without resolving nest-tube-level entry and exit events. C_LIO_LIWe developed BeeMonitor, an integrated hardware and computer-vision pipeline that detects nest entry and exit events in cavity-nesting solitary bees from continuous video, using Osmia cornifrons (the horn-faced mason bee) as a model system. A low-cost Raspberry Pi handles solar-powered field recording, while the software combines object detection (YOLOv26), a custom multiple-object tracker (BeeTrack), and a Random Forest classifier trained on trajectory-derived features to distinguish genuine events from incidental detections. C_LIO_LIOver a 29-day deployment, hardware reliability averaged 97.5% recording coverage. The pipeline achieved 91.3% precision and 87.3% recall (F1 = 0.893), generalizing robustly under leave-one-video-out cross-validation (mean F1 = 0.904). Detected foraging trips correlated strongly with brood cell counts (R2 = 0.849, p < 0.001, n = 19), and a Random Forest model (AUC = 0.820) identified solar radiation as the dominant driver of foraging activity, followed by temperature. C_LIO_LIBeeMonitor demonstrates that automated computer vision can reliably extract ecologically relevant behavioral data from continuous video, enabling real-time analysis of pollinator behavior and abundance at a temporal and spatial resolution unattainable through manual observation. Its modular design supports adaptation to other species and monitoring contexts. C_LI

O_LIMonitoring and limiting the spread of invasive species on islands requires efficient detection and population estimation methods. However, elusive species can be difficult to monitor using traditional methods, making autonomous approaches such as camera trapping and acoustic monitoring increasingly valuable. C_LIO_LIOn the island of Okinawa, Japan, the small Indian mongoose ( Urva auropunctata) threatens many native species since its introduction in 1910. Listed among the worlds worst invasive species, effective monitoring of U. auropunctata in Okinawa is critical. The Okinawa Environmental Observation Network (OKEON) uses camera traps to detect U. auropunctata, but success depends on precise placement. Though OKEON also includes a high-resolution acoustic monitoring programme, no audio classification model currently exists for U. auropunctata. Developing such a model could improve substantially our capacity to detect and manage the species. C_LIO_LIUsing sparse U. auropunctata vocalisations collected from camera trap videos, we built a lightweight Convolutional Neural Network distilled from a more complex model for classifying contact calls and alarm calls of U. auropunctata. Our distilled model performed similarly to the full model at detecting vocalisations from training data, but was considerably faster. C_LIO_LIWe applied the distilled classifier to [~]486 hrs of audio collected over eight years from southern Okinawa, where we successfully detected U. auropunctata a handful of times in each year of recording. In spite of strong model performance on test data, our model did not transfer well to unseen data, perhaps owing to the rarity of U. auropunctata calls and consequent small training dataset size, limiting its utility for ecological monitoring. C_LIO_LIPractical implication. The use of sparse audio data from camera trap videos to train an acoustic classifier had limited utility to detect the rarely vocalising U. auropunctata from passive acoustic monitoring data. We provide several recommendations for enhancing classifier performance to provide robust actionable insights into the distribution and spread of U. auropunctata, and aid targeted conservation efforts for Okinawas threatened biodiversity. C_LI

0.O_LIMorphological traits such as floral area and body size are fundamental to ecological research, serving as inputs for studies of pollinator-plant interactions, habitat quality, and biodiversity monitoring. However, accurately measuring these traits from images remains challenging, particularly in complex field conditions where existing tools exhibit reduced accuracy and limited generalizability across taxa. C_LIO_LIWe present EcoMorph, a modular morphological measurement system that leverages the Segment Anything Model 3 (SAM3) to quantify traits across diverse ecological contexts. Unlike task-specific segmentation models requiring domain-specific training data, SAM3s prompt-based architecture enables segmentation of arbitrary biological structures from natural-language prompts, using the same underlying model across flowers, insects, and other targets without retraining. From the resulting segmentations, EcoMorph extracts three classes of measurement: area, linear dimensions, and object counts. C_LIO_LIWe validated EcoMorph across two ecological scales. At the intermediate scale, EcoMorph-derived floral area agreed closely with manual ImageJ measurements (R2 = 0.935, n = 74) under simple-background conditions and (R2 = 0.928, n = 58) under complex-background conditions, with valid predictions for 95% of images. At the fine scale, EcoMorph-derived insect body area was strongly correlated with hand-measured intertegular distance (r = 0.810, n = 349), capturing body-size variation across species from the small Bombus impatiens to the large Xylocopa virginica. Object counts matched manual counts almost exactly for well-separated insects in an insect box (R2 = 0.9997, n = 12). C_LIO_LIBy combining prompt-based segmentation with modular measurement, EcoMorph enables high-throughput quantification of area, size, and abundance from heterogeneous image sources without taxon-specific training. This generality supports a broad range of ecological applications, including pollinator and plant trait research, biodiversity and abundance monitoring, and allometric biomass estimation. C_LI