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.

Tropical forest canopies contain many tree and liana species, and foliar and reproductive characteristics useful for taxonomic identification are often difficult to see from the forest floor. As such, taxonomic identification often becomes a bottleneck in tropical forest inventories. Here we present a drone-based workflow to automatically acquire large volumes of close-up, ultra-high resolution photos of selected tree crowns (or specific locations over the canopy) to support tropical botanical and ecological studies (https://youtu.be/80goMEifpc4). Our workflow is built around the small, easy-to-use DJI Mavic 3 Enterprise (M3E) drone, which is equipped with a wide-angle and a telephoto camera. On day one, the pilot maps a forest area of up to [~]200 ha with the wide-angle camera to generate a high-resolution digital surface model (DSM) and orthomosaic using structure-from-motion (SfM) photogrammetry. On subsequent days, the pilot acquires close-up photos with the telephoto camera from up to 300 selected canopy trees per day. These close-up photos are acquired from 6 m above the canopy and contain a high level of visual detail that allows botanists to reliably identify many tree and liana species. The photos are geolocated with survey-grade accuracy using RTK GNSS, thus facilitating spatial co-registration with other data sources, including the photogrammetry products. The primary operational challenge of our workflow is the need to maintain RTK corrections with the drone to ensure that close-up photos are acquired exactly at the predefined locations. The maximum operational range we achieved was 3 km, which would allow the pilot to reach any tree within a [~]2800 ha area from the take-off point. Although our workflow was developed to support taxonomic identification of tropical trees and lianas, it could be extended to any other forest or vegetation type to support botanical, phenological, and ecological studies. We provide harpia, an open-source Python library to program these automatic close-up photo missions with the M3E drone (https://github.com/traitlab/harpia). Data/code for peer review statementWe provide harpia, an open-source Python library to program these automatic close-up photo missions (https://github.com/traitlab/harpia). Drone imagery and labelled close-up photo data are not yet publicly available because they were acquired with the goal of publishing benchmark machine learning datasets and models for tree and liana species classification and prior publication of the data would jeopardize this future publication.

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.

Unmanned aerial vehicles (UAVs) are increasingly used for high throughput phenotyping. In principle, freely flown vehicles would permit real-time flexibility in identifying and scouting regions of interest. Mosaicking multiple images provides a high resolution global image and consumer-grade UAVs offer low cost, ease of flying, and excellent RGB cameras. The vehicles inaccurate telemetry complicates estimating the homographies between pairs of frames, the standard mosaicking approach. Moreover, errors accumulate during computation, distorting later portions of the mosaic. Finally, crop fields are particularly challenging to mosaic because their planting is so regular and the plants are so similar, eliminating distinctive features that could guide mosaicking. We propose MaiZaic, an end-to-end pipeline that dynamically samples video frames using optical flow, automates camera and gimbal calibration, estimates homographies with an unsupervised convolutional neural network, detects shots among frames, and generates mini-mosaics. Together, these techniques significantly reduce errors in the output mosaics. Our deep learning model is trained on a comprehensive video dataset comprising different flight trajectories, maize lines, growth stages, and augmented illumination data. MaiZaic is more accurate and faster than ASIFT and more robust than our earlier CorNet and CorNetv2. We demonstrate MaiZaics effectiveness in generating accurate mosaics of imagery captured by freely-flown UAVs and explore its generalizability. Core ideasO_LIMaiZaic is an end-to-end pipeline to mosaic freely flown agricultural imagery captured with consumer-grade UAVs. C_LIO_LIMaiZaic introduces novel algorithms that efficiently choose video frames, calibrate, and mosaic the imagery. C_LIO_LIOur unsupervised deep homography estimator, CorNetv3, is 14 times faster and 8.59% more accurare than ASIFT. C_LIO_LIMaiZaic generalizes well and mosaicks maize at different growth stages, objects, trajectories, cameras, and pilots. C_LIO_LIThe mini-mosaicking algorithm improves mosaic accuracy by interrupting error accumulation. C_LI

Thermal drones offer significant advantages for monitoring wildlife in low-light conditions; however, detection performance is influenced by technical settings and environmental factors. This study evaluates the use of nocturnal drone surveys to detect macropods in Tasmania, with the aim of determining optimal flight parameters for balancing detection accuracy and survey efficiency. Field surveys were conducted in Narawntapu National Park using varying flight altitudes and image overlap settings, and detection rates were compared across thermal datasets. Manual annotations and basic thresholding method were used to quantify detection success. Results showed that detection rates were highest when surveys were conducted under cooler ambient temperatures and at moderate altitudes (e.g. 60 metres AGL) with 50% image overlap. These findings provide practical guidance for designing nocturnal drone surveys and offers baseline recommendations for using drone-mounted thermal sensors to monitor large-bodied, crepuscular mammals, with broader implications for scalable wildlife monitoring programs.

O_LIThe exodus of swallows from communal nighttime roosts is often visible as an expanding ring-shaped pattern in weather radar data. The WSR-88D network operated by the National Weather Service archives more than 25 years of data across 143 stations in the contiguous US. However, access to information about the roosting behavior of swallows is limited by the cost of manual annotation of these scans. C_LIO_LIWe develop an AI system to detect and track swallow roosts in weather radar data. Our model is based on the Faster R-CNN architecture and is customized to incorporate multiple spatial and temporal channels in volumetric radar scans using novel adaptor layers. We systematically study the impact of network architecture and input representation for this task. We incorporate our detection outputs into an AI-assisted system with an interface for human screening to collect research-grade data about roosting behavior. We deploy the system to collect information from 12 radar stations in the Great Lakes region of the US spanning 21 years. C_LIO_LIThe addition of temporal information improves roost detection performance from 47.0% mean average precision to 54.7%. Temporal information helps the model recognize the expanding pattern of roosts and filter false positives due to rain and static structures. Our system allowed the annotation of 15,628 roost signatures with 64,620 single-frame detections in 612,786 radar scans with 183.6 total hours of human screening, or 1.08 seconds per radar scan. C_LIO_LIOur AI-assisted system provides research-quality roost data with far less human effort than manual annotation of radar scans. The data contains critical information about the phenology and population trends of swallows and martins, a declining group of aerial insectivores. Our successful deployment to collect historical data for 8% of the radar stations in the contiguous US lays the groundwork for continentscale analysis of swallow roosts, and provides a starting point for analysis of other family-specific phenomena in weather radar, such as bat roosts and mayfly hatches. C_LI

Manual annotation of drone imagery is labour-intensive and prone to observer bias, particularly when applied to large datasets across varied environments. To address this, a deep-learning pipeline was developed and evaluated for identifying macropods in visual (RGB) drone imagery, using a convolutional neural network (CNN) adapted from the DeepForest framework. The model was trained on annotated images of Forester kangaroos (Macropus giganteus tasmaniensis) and Bennetts wallabies (Notamacropus rufogriseus) collected across two Tasmanian study sites. Performance was assessed using independent test sets from each site, representing open and forest-edge habitats, as well as a combined multi-site test set. Detection accuracy was quantified using precision, recall, and F1 scores, with further analyses evaluating the effect of solar altitude angle on model performance. The model achieved high recall across sites, indicating strong potential for minimising missed detections under diverse conditions. These results demonstrate the feasibility of applying transfer learning to drone-based wildlife surveys and highlight the promise of deep learning models for reducing manual effort in macropod monitoring, with applications for broader conservation and management workflows.

Accurate monitoring of populations is essential for conservation management, including for vulnerable seabirds. Yet traditional ground-based surveys are logistically challenging and time-consuming, especially in remote environments such as the sub-Antarctic islands. Advances in aerial imagery and artificial intelligence (AI) offer opportunities to improve the efficiency and repeatability of population surveys. In this study, we evaluate an AI-based approach for counting Salvins albatross from high-resolution aerial imagery collected using a piloted fixed-wing aircraft at the Bounty Islands, New Zealand. Imagery acquired during a single-day survey was processed to create orthomosaic images, which were previously analysed using manual counts by an experienced observer. We applied an automated detection and counting model based on a Faster R-CNN architecture with Slicing-Aided Hyper-Inference, and compared AI-derived counts with original human counts in terms of accuracy, consistency, and processing time. The AI achieved an initial F1 score of 92.8% for albatross detection and produced counts within 3% of the manual results, while reducing processing time from approximately 66 hours to just over four minutes. The model was also capable of simultaneously detecting additional species present within the mixed breeding colony, including erect-crested penguins, fulmar prions, and New Zealand fur seals, adding scalable efficiency gains for future surveys. Our results demonstrate that combining piloted aircraft surveys with AI-based image analysis provides a rapid, scalable, and accurate method for monitoring seabird populations, with substantial benefits for conservation management in remote and logistically constrained regions.

Global environmental challenges require comprehensive data to manage and protect biodiversity. Currently, vision-based biodiversity monitoring efforts are mixed, incomplete, human-dependent, and passive. To tackle these issues, we present a portable, modular, low-power device with embedded vision for biodiversity monitoring. Our camera uses interchangeable lenses to resolve barely visible and remote subjects, as well as customisable algorithms for blob detection, region-of-interest classification, and object detection to identify targets. We showcase our system in six case studies from the ethology, landscape ecology, agronomy, pollination ecology, conservation biology, and phenology disciplines. Using the same devices, we discovered bats feeding on durian tree flowers, monitored flying bats and their insect prey, identified nocturnal insect pests in paddy fields, detected bees visiting rapeseed crop flowers, triggered real-time alerts for waterbirds, and tracked flower phenology over months. We measured classification accuracies between 55% and 96% in our field surveys and used them to standardise observations over highly-resolved time scales. The cameras are amenable to situations where automated vision-based monitoring is required off the grid, in natural and agricultural ecosystems, and in particular for quantifying species interactions. Embedded vision devices such as this will help addressing global biodiversity challenges and facilitate a technology-aided global food systems transformation.

O_LICamera trap arrays can generate thousands to millions of images that require exorbitant time and effort to classify and annotate by trained observers. Computer vision has evolved as an automated alternative to manual classification. The most popular computer vision solution is the supervised Machine Learning technique, which uses labeled images to train automated classification algorithms. C_LIO_LIWe propose a multi-step semi-automated workflow that consists of (1) identifying and separating bad-from good-quality images, (2) parsing good images into animals, humans, vehicles, and empty, and (3) cropping animals from images and classifying them into species for manual inspection. We trained, validated, and evaluated this approach using 548,627 images from 46 cameras in two regions of the Arctic (northeastern Norway, and Yamal Peninsula, Russia). C_LIO_LIWe obtained an accuracy of 0.959 for all three steps combined with the complete year test data set at Varanger and 0.922 at Yamal, reducing the number of images that required manual inspection to 7.9% of the original set from Varanger and 3.2% from Yamal. C_LIO_LIResearchers can modify this multi-step process to meet their specific needs for monitoring and surveying wildlife, providing greater flexibility than current options available for image classification. C_LI

Tree detection is a fundamental task in remote sensing for forestry and ecosystem ecology applications. While many individual tree segmentation algorithms have been proposed, the development and testing of these algorithms is typically site specific, with few methods evaluated against data from multiple forest types simultaneously. This makes it difficult to determine the generalization of proposed approaches, and limits tree detection at broad scales. Using data from the National Ecological Observatory Network we extend a recently developed semi-supervised deep learning algorithm to include data from a range of forest types, determine whether information from one forest can be used for tree detection in other forests, and explore the potential for building a universal tree detection algorithm. We find that the deep learning approach works well for overstory tree detection across forest conditions, outperforming conventional LIDAR-only methods in all forest types. Performance was best in open oak woodlands and worst in alpine forests. When models were fit to one forest type and used to predict another, performance generally decreased, with better performance when forests were more similar in structure. However, when models were pretrained on data from other sites and then fine-tuned using a small amount of hand-labeled data from the evaluation site, they performed similarly to local site models. Most importantly, a universal model fit to data from all sites simultaneously performed as well or better than individual models trained for each local site. This result suggests that RGB tree detection models that can be applied to a wide array of forest types at broad scales should be possible.

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.

1. Camera traps have become a cornerstone of wildlife ecological research, yet the manual analysis of the millions of images they generate requires substantial time and resources. Deep learning-based automation has emerged as a promising solution, existing global general-purpose models exhibit limitations in precisely recognizing local endemic species and adapting to unique local ecosystems. 2. This study developed a high-performance classification model optimized for native species. A large-scale "Korean Wildlife Dataset" was constructed from data collected across diverse domestic habitats, and a novel architecture was proposed to overcome limitations of conventional CNNs. The proposed Graph Attention Contrastive Learning (GACL) model is structured as a two-stage pipeline. Stage one employs YOLOv5 and MegaDetector to detect animals, humans, and vehicles, filtering valid images. Stage two performs fine-grained species classification. GACL captures structural relationships among object parts using a Graph Attention Transformer (GAT) and aligns semantic correspondence between images and textual descriptions via Parallel Contrastive Learning, enabling deeper understanding beyond simple visual features. 3. Evaluation on an independent test set demonstrated that the proposed model robust classification performance with an overall accuracy of 96.83% across four classes (Wildboar, Goral, Deers, and Other). Notably, in a comparative analysis against a global general-purpose model, our model showed distinct advantages in the precise recognition of endemic species. Furthermore, it exhibited a lower false positive rate in identifying animals in empty images, confirming its potential to enhance the efficiency of the data cleaning process. 4. Beyond technical accuracy, this study highlights that region-specific AI models that reflect local ecological characteristics can provide substantial practical value for wildlife monitoring and biodiversity conservation. Future work will require continuous efforts in data diversification and model lightweighting to further improve model robustness and practicality.

O_LIA time-consuming challenge faced by camera trap practitioners all over the world is the extraction of meaningful data from images to inform ecological management. The primary methods of image processing used by practitioners includes manual analysis and citizen science. An increasingly popular alternative is automated image classification software. However, most automated solutions are not sufficiently robust to be deployed on a large scale. Key challenges include limited access to images for each species and lack of location invariance when transferring models between sites. This prevents optimal use of ecological data and results in significant expenditure of time and resources to annotate and retrain deep learning models. C_LIO_LIIn this study, we aimed to (a) assess the value of publicly available non-iconic FlickR images in the training of deep learning models for camera trap object detection, (b) develop an out-of-the-box location invariant automated camera trap image processing solution for ecologist using deep transfer learning and (c) explore the use of small subsets of camera trap images in optimisation of a FlickR trained deep learning model for high precision ecological object detection. C_LIO_LIWe collected and annotated a dataset of images of "pigs" (Sus scrofa and Phacochoerus africanus) from the consumer image sharing website FlickR. These images were used to achieve transfer learning using a RetinaNet model in the task of object detection. We compared the performance of this model to the performance of models trained on combinations of camera trap images obtained from five different projects, each characterised by 5 different geographical regions. Furthermore, we explored optimisation of the FlickR model via infusion of small subsets of camera trap images to increase robustness in difficult images. C_LIO_LIIn most cases, the mean Average Precision (mAP) of the FlickR trained model when tested on out of sample camera trap sites (67.21-91.92%) was significantly higher than the mAP achieved by models trained on only one geographical location (4.42-90.8%) and rivalled the mAP of models trained on mixed camera trap datasets (68.96-92.75%). The infusion of camera trap images into the FlickR training further improved AP by 5.10-22.32% to 83.60-97.02%. C_LIO_LIEcology researchers can use FlickR images in the training of automated deep learning solutions for camera trap image processing to significantly reduce time and resource expenditure by allowing the development of location invariant, highly robust out-of-the-box solutions. This would allow AI technologies to be deployed on a large scale in ecological applications. C_LI

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.

O_LIThe ability of drones to provide detailed information on animals and their surroundings makes them ideal for studying animal behaviour at fine scales. While drones can provide high-resolution images of what animals are doing, they should also, in theory, be able to provide data on where they are. However, reconstructing geo-referenced tracks from drone videos that follow animals is challenging, particularly because current methods require specific drone flight patterns and large computational power. C_LIO_LIHere, we combine deep learning and object tracking methods with a novel geo-referencing algorithm which allows us to track individuals across video frames and reconstruct their geo-referenced trajectories. We used a Region-based Convoluted Neural Network to detect animals and a Hungarian tracking algorithm to link detections across video frames, and then geo-referenced each detection in every frame to reconstruct individual trajectories. C_LIO_LIWe tested our geo-refencing algorithm through multiple drone flights with varying flight parameters over known Ground Control Points. The median (95% CI) geo-referencing error was 2.81 (0.74 - 23.23) meters, which reduced by 50% when the drone camera was positioned between -90{degrees} and -40{degrees}. Error increased with drone height and camera angle (-90{degrees} refers to the camera pointing towards the ground) but was not impacted by drone orientation. C_LIO_LIWe then demonstrate the utility of our framework with empirical examples using consumer-level drones. First, we tracked a volunteer carrying a high-resolution GPS unit and overlayed their GPS tracks on our estimated tracks to quantify tracking error. Next, we used drone videos of two delphinid species (Tursiops truncates gephyreus, and Sousa plumbea) representing varying environmental and flight conditions. We were able to successfully infer individual tracks across all conditions except when individuals formed tight clusters, in which case tracks were assigned a group identifier. C_LIO_LIOur framework demonstrates an easy and robust approach to translate drone videos of moving animals into geo-referenced animal tracks which is applicable in many research contexts. A major advance over previous methods is that our algorithm is robust to different camera angles, and provides tracks with accuracy on-par with, or even exceeding, the accuracy from GPS tracking. C_LI HeadlineAutomated drone-based animal tracking

Developing actionable early detection and warning systems for agricultural stakeholders is crucial to reduce the annual $200B USD losses and environmental impacts associated with crop diseases. Agricultural stakeholders primarily rely on labor-intensive, expensive scouting and molecular testing to detect disease. Spectroscopic imagery (SI) can improve plant disease management by offering decision-makers accurate risk maps derived from Machine Learning (ML) models. However, training and deploying ML requires significant computation and storage capabilities. This challenge will become even greater as global scale data from the forthcoming Surface Biology & Geology (SBG) satellite becomes available. This work presents a cloud-hosted architecture to streamline plant disease detection with SI from NASAs AVIRIS-NG platform, using grapevine leafroll associated virus complex 3 (GLRaV-3) as a model system. Here, we showcase a pipeline for processing SI to produce plant disease detection models and demonstrate that the underlying principles of a cloud-based disease detection system easily accommodate model improvements and shifting data modalities. Our goal is to make the insights derived from SI available to agricultural stakeholders via a platform designed with their needs and values in mind. The key outcome of this work is an innovative, responsive system foundation that can empower agricultural stakeholders to make data-driven plant disease management decisions, while serving as a framework for others pursuing use-inspired application development for agriculture to follow that ensures social impact and reproducibility while preserving stakeholder privacy. Key PointsO_LICloud-based plant disease detection system, easily accommodates newly developed and/or improved models, as well as diverse data modalities. C_LIO_LIEmpower agricultural stakeholders to use hyperspectral data for decision support while preserving stakeholder data privacy. C_LIO_LIOutline framework for researchers interested in designing geospatial/remote sensing applications for agricultural stakeholders to follow. C_LI

Historical remote sensing imagery, including digital Landsat and analog Keyhole have potential to inform Antarctic conservation by providing insights into habitat and population dynamics. A case in point is that their long temporal range, extending as far back as the 1960s, offers an unparalleled look into the distribution of Emperor Penguin (Aptenodytes forsteri) colonies and their habitat. Here we demonstrate that Keyhole and Landsat sensors are capable of detecting presence and long-term change in penguin guano on sea-ice despite challenging environmental conditions. For 18 of the 66 known emperor penguin colonies, we confirmed their presence in images that predate earliest published records. We further used dense time series from 1960s to 2024 of the Cape Washington colony to exemplify change in guano area over time - but found little variation in guano area. We show that the guano area undergoes two distinct stages during a breeding season (a stable and clustered early stage, and a diffused and widespread late stage). We correlated guano area with penguin colony size and found that guano area correlates with colony size (Spearmans {rho} = 0.59, p-value = 0.017). Taken together, our results suggest long term resilience of penguin colonies in face of global change, and highlight that the use of historical remote sensing imagery across the Antarctic can inform conservation efforts and benefit the ongoing historical studies of penguin colony dynamics. We highlight the benefit of remote sensing for documenting otherwise inaccessible colonies and for quantifying and understanding historical colony dynamics.

O_LIDemographic data are essential to construct mechanistic models to understand how populations change over time and in response to global threats like climate change. Existing demographic data are either lacking or insufficient for many species, particularly those that are challenging to study, such as marine mammals. A pipeline for collecting accurate demographic data to construct robust demographic models at scale would fill this knowledge gap for many species, including marine mammals like pinnipeds (seals, sea lions, and walruses). C_LIO_LIWe introduce a non-invasive pipeline to estimate the 3D body size (volume) of species that will allow monitoring at high spatial and temporal scales. Our pipeline integrates 3D structure-from-motion photogrammetry data collected via planned flight missions using off-the-shelf, multirotor unmanned aerial vehicles (UAVs). We apply and validate this pipeline on the grey seal Halichoerus grypus, a marine species that spends much of its time at sea but is predictably observable during its annual breeding season. We investigate the optimal ground sampling distance (GSD) for surveys by calculating the success rates and accuracy of volume estimates of individuals at different elevations. C_LIO_LIWe establish an optimal GSD of 0.8 cm px-1 for animals similar in size to UK grey seals ([~]1.4 - 2.5 m length), making our pipeline reproducible and applicable to a broad range of organisms. Volume estimates were accurate and could be made for up to 68% of hauled-out seals in the study areas. Finally, we highlight six key traits that make a species well-suited to estimating body volume following this pipeline. Good candidates include large reptiles like crocodiles, large mammals such as hippopotamus, and shrubs or bushes in deserts and Mediterranean habitats. C_LIO_LIOur pipeline accurately estimates individual body volume of marine macrovertebrates in a time-and cost-effective manner whilst minimising disturbance. Whilst the approach is applied to pinnipeds here, the pipeline is adaptable to many different taxa that are otherwise challenging to study. Our proposed approach therefore opens up previously inaccessible areas of the Tree of Life to demographic studies, which will improve our ability to protect and conserve these species into the future. C_LI