Substantial progress is expected in the coming decade as next-generation Earth Observation missions, multi-sensor integration, and algorithm development converge. These advances promise more precise measurements of essential biodiversity variables such as ecosystem extent, structure and condition. Yet key challenges remain in translating electromagnetic signals into biologically meaningful metrics, scaling from field plots to global extents, integrating multi-source datasets while accounting for uncertainty, and aligning products with ecological theory, conservation practice, and global policy frameworks.
This session invites contributions that showcase how novel remote sensing and AI methods support biodiversity research and conservation. We particularly encourage studies that link remote sensing with in situ data, develop scalable approaches, advance ecological modelling to predict biodiversity change and its drivers, and demonstrate monitoring frameworks combining remote sensing, in situ networks and novel methods. By uniting advances in sensing technology and biodiversity science, the session will highlight how remote sensing can help contribute to the Kunming-Montreal Global Biodiversity Framework, the SDGs, and other international targets.
Discover how cutting-edge remote sensing and AI are transforming biodiversity science—connecting technology, field data, and global conservation goals to drive smarter, scalable action for our planet.
Biodiversity monitoring is important to support decision-making for managing landscapes sustainably and supporting national and international policy targets for nature conservation and restoration, including the Global Biodiversity Framework. While assessing the status, change and drivers of biodiversity remains challenging, we have new opportunities to support biodiversity monitoring from space with a growing suite of Earth observation satellites. Remote sensing is especially well suited to monitor ecosystems in terms of their vegetation structure and forest structural diversity with lidar and radar, as well as vegetation functions, foliar functional diversity and community composition with imaging spectroscopy and multispectral imaging. In this talk, we will provide examples for monitoring forest structural diversity using the spaceborne lidar GEDI. We evaluated and compared structural diversity across contrasting biomes in the Western US and Central Africa, and we found that general biogeographic patterns of higher horizontal structural diversity in areas with higher disturbance, higher topographic variation and lower aridity hold across continents and scales. For monitoring ecosystem functions, we will provide examples for monitoring plant traits and functional diversity using imaging spectroscopy along elevation gradients in California. We will provide insights into the role of trait-trait relationships and trait selection for mapping trait diversity patterns at the landscape scale. We found that diversity patterns vary by the type and number of functional traits included in the analyses, and that the interpretation is context dependent. And for monitoring composition, we will provide examples indicating how well we can distinguish different vegetation types, communities and species with spectroscopy, and how well we predict animal composition and niche space using a remote sensing-based biodiversity data cube, BioCube. With these examples, we will demonstrate new capabilities and avenues for monitoring different aspects of biodiversity change using remote sensing at the landscape scale, and we will provide important context for the interpretation of these results. Remote sensing can provide information about biological communities and habitats, ecosystems and biomes at different spatial scales and time steps that should be integrated with other biodiversity data, models and decision support tools to fully leverage its potential for biodiversity monitoring.
How to cite: Schneider, F. D., Zheng, T., Ferraz, A., Berman, L., Jakobsen, C. D., Revenga, J. C., Wu, Z., Ye, Z., Pavlick, R. P., Townsend, P. A., and Normand, S.: Monitoring Ecosystem Structure, Function and Composition with Remote Sensing, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-288, https://doi.org/10.5194/wbf2026-288, 2026.
Imaging spectroscopy is a versatile technology for acquiring spectral information about the Earth’s surface across ecosystems, typically covering the wavelength range of the electromagnetic spectrum between 400 and 2500 nm. Imaging spectrometers can be installed on varying platforms (e.g., drones, aircrafts, satellites) and provide valuable downstream products for biodiversity monitoring and understanding across multiple spatial scales. Exemplary downstream products for grasslands include species composition, plant life-forms, plant traits, and indicator values.
Due to the reflectance anisotropy of natural surfaces, i.e., the non-uniform scattering of incident light from surfaces in different directions, the signal measured by a spectrometer may still exhibit surface type specific dependencies on the observation and illumination geometry that also vary across wavelengths. Without adequate compensation for these effects, downstream products can be impacted, with consequences for change detection, causal inference, and management strategies. Although several compensation methods have been developed, they are not yet incorporated into most data processing pipelines, and our understanding of their effectiveness on various downstream products across ecosystems remains limited.
Using an airborne imaging spectroscopy dataset acquired over the Swiss National Park, we aim to quantify and analyze the effect of applying a reflectance anisotropy compensation method when studying alpine grassland ecosystems. Our focus lies on grassland canopy trait information derived from acquired spectroscopy data using the PROSAIL radiative transfer model. We particularly focus on the canopy traits leaf area index, chlorophyll content, and leaf mass per area that are often used to quantify the diversity, functioning and resilience of vegetation ecosystems. We investigate the capacity of a reflectance anisotropy compensation method to reduce anisotropy effects in data covering the topographically challenging Swiss National Park, and to eventually improve the spatial consistency of derived grassland canopy traits.
Our results facilitate multi-temporal biodiversity assessments based on spatially consistent grassland canopy trait information derived from imaging spectroscopy data. Insights are also relevant for validation attempts of ongoing and future spaceborne imaging spectroscopy missions. Furthermore, we critically reflect on the current state and future needs of methods available to compensate for reflectance anisotropy.
How to cite: Schweizer, J., Rossi, C., Damm, A., Schläpfer, D., Hüni, A., Ginzler, C., Wegner, J. D., and Kneubühler, M.: Improving spatial consistency of grassland canopy traits derived from imaging spectroscopy data through reflectance anisotropy compensation, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-313, https://doi.org/10.5194/wbf2026-313, 2026.
Recent advances in remote sensing technology and methods have generated an unprecedented number of biodiversity-related products, significantly enhancing our ability to monitor and understand biodiversity. The combination of high-revisit satellites, such as harmonized Landsat and Sentinel-2 (HLS), along with spaceborne LiDAR data from the Global Ecosystem Dynamics Investigation (GEDI), provides both temporal frequency and structural information at unprecedented detail. Numerous datasets are now accessible, either as analysis-ready products or variables that can be generated from the growing archive of satellite observations. These data include topographical features, land cover, land use, phenology, productivity, plant traits and their diversity. However, these observations and products are scattered across different platforms, making data discovery, preparation, and processing laborious and time-consuming. To address this limitation, we compiled and standardized a collection of over 70 remote sensing products relevant to biodiversity - including vegetation structure, plant traits, land cover classes, dynamic habitat indices, spectral diversity and habitat heterogeneity. All datasets have a very high spatial resolution (10–100 m) and cover the entire Alpine Region (as delineated by the European Union Strategy for the Alpine Region - EUSALP). By integrating these products with in-situ biodiversity monitoring records from Switzerland for birds, butterflies, and plants, we demonstrate how remote sensing data can enable landscape-scale biodiversity modelling to identify biodiversity hotspots and assess the representativeness of monitoring networks across diverse Alpine ecosystems. These models achieved robust predictions of species richness at the landscape scale (1 km²), with cross-validated R² values exceeding 0.7 across the three taxonomic groups. Independent validation using monitoring data from Germany, Austria, and Italy further confirmed the potential of remote sensing datasets for developing accurate and transferable modelling approaches. The entire dataset will be made openly available to facilitate the integration of remote sensing data into species distribution and macroecological models, providing improved potential for both prediction and ecological inference.
How to cite: Rossi, C., Dörig, F., Corsini, M., Frühholz, K., Tello-García, E., Hauser, L. T., König, S., Leitinger, G. F., Marsoner, T., and Rüdisser, J. M.: From space to species: A high-resolution dataset of biodiversity-relevant remote sensing products for the Alps , World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-325, https://doi.org/10.5194/wbf2026-325, 2026.
Previous studies have shown that multitemporal data can strengthen the relationship between grassland diversity and spectral reflectance. However, most studies used interpolated Sentinel-2 time series as such. This study investigated whether species richness models can be improved using fitted Sentinel-2 temporal features describing the overall plant growth patterns.
We measured plant species richness in 77 quadrats (1m x 1m) from revegetated alpine grasslands located around an open pit mine in Hochfilzen, Austria, characterized by a high range of species richness. Several multi-year Sentinel-2 vegetation index time series were used as inputs for fitting temporal features such as phenology descriptors, harmonic decomposition, frequency decomposition and functional principal components. Those features were compared to interpolated time series of Sentinel-2 bands used in state of the art baseline models (Fauvel et al., 2020; Muro et al., 2022).
The feature sets were inserted into a Random Forest regression model pipeline, first selecting the best performing features in a nested cross-validation, then applying the final model over the grassland areas to produce species richness maps. Lastly, SHAP feature analysis was performed to improve model interpretability.
Our best model, using fitted CIRE time series features, achieved coefficient of determination R2 = 0.19 in cross-validation and R2 = 0.36 on holdout set (16 quadrats). Features describing events around peak growing season were found especially important. Our model clearly outperformed all baseline models on holdout set across all metrics: R2 (+0.15 to +0.33), absolute Root Mean Squar Error RMSE (-0.35 to -0.91) and relative RMSE (-0.02 to -0.05).
All models highlighted similar areas of high or low richness. Differences were observed for most extreme species richness, or less densely vegetated pixels. Results suggest our features are better suited to small datasets. The comparison of models will also be carried out on larger field inventories.
Future steps will include extension to species abundance metrics, and a comparison to spectral variation features extracted from multi-scale hyperspectral imagery from Hyspex Mjölnir VS-620 drone and EnMAP satellite.
This research is part of the MultiMiner project funded by the European Union’s Horizon Europe research and innovations actions programme under Grant Agreement No. 101091374.
How to cite: Lindgren, P., Molinier, M., Kiessling, A., and Tischler, A.: Improving plant diversity prediction in revegetated grasslands using compact Sentinel-2 time series descriptors, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-342, https://doi.org/10.5194/wbf2026-342, 2026.
Occupancy models are tools for modeling the relationship between habitat and species occurrence while accounting for the fact that species may still be present even if not detected. The types of environmental variables typically used for characterizing habitats in such ecological models, such as precipitation or tree cover, are frequently of low spatial resolution, with a single value for a spatial pixel size of, e.g., 1 km2. This spatial scale fails to capture the nuances of micro-habitat conditions that can strongly influence species presence, and additionally, as many of these are derived from satellite data, there are aspects of the environment they cannot capture, such as the structure of vegetation below the forest canopy. To address these gaps, we propose to combine high-resolution satellite and ground-level imagery to produce multi-modal environmental features that better capture micro-habitat conditions, and incorporate these multi-modal features into hierarchical Bayesian species occupancy models. We leverage pre-trained deep learning models to flexibly capture relevant information directly from raw imagery, in contrast to traditional approaches which rely on derived and/or hand-crafted sets of ecosystem covariates. We implement deep multi-modal species occupancy modeling using a new open-source Python package for ecological modeling, designed for bridging machine learning and statistical ecology. We test our method under a strict evaluation protocol on various mammal species across thousands of camera traps in Snapshot USA surveys, and find that multi-modal features substantially enhance predictive power compared to traditional environmental variables alone. To aid in interpreting these models, we propose a technique based on vision-language models that automatically extracts habitat elements that are particularly influential on model estimates. Our results not only highlight the predictive value and complementarity of satellite and in-situ imagery, but also make the case for more closely integrating deep learning models and traditional statistical ecological models while maintaining their interpretability.
How to cite: Haucke, T., Harrell, L., Shen, Y., Klein, L., Rolnick, D., Gillespie, L. E., and Beery, S.: Deep Multi-modal Species Occupancy Modeling, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-347, https://doi.org/10.5194/wbf2026-347, 2026.
Numerous studies have shown the immense potential of Sentinel-2 satellite data for tree species classification across temperate forest ecosystems. However, key challenges such as the classification of minor tree species and detection of admixture species remain. In particular, when mapping tree species classes at a national scale, heterogeneous landscapes, different forest management practices and differences in phenology introduce additional uncertainties. Nevertheless, detailed information on tree species composition, including minor species, is crucial for nature conservation, biodiversity monitoring, and the development of forest management strategies opting for enhanced resilience. At the same time, spatially explicit information on model uncertainties and domains are often lacking in state-of-art national tree species maps. Here, we aim to provide new insight in mapping minor tree species relevant for nature conservation and forest resilience at the landscape scale in Germany by combining diverse reference data sources such as forest management maps, biotope mapping data, and regional forest inventories with Sentinel-2 timeseries and periodic forest masks. The obtained tree species maps are complemented by spatially explicit uncertainty maps and areas of model applicability, allowing for a transparent and species-specific assessment of the classification results. Subsequently, these outputs will be used to derive tree species composition and tree species diversity measures across different spatial scales. We compare our findings to existing tree species maps and aim to quantify tree species diversity measures in relation to in-situ inventory plots across different landscape regions in Germany. Additionally, the generation of periodic tree species maps enables the analysis of temporal changes and shifts in tree species composition and biodiversity, especially in view of the drastic changes and stresses that forests in Germany have been exposed to in recent years as a result of climate change. This study contributes to a more comprehensive understanding of the potential and the limitations of remotely sensed tree species maps for monitoring biodiversity and supporting conservation planning across spatial and temporal scales.
How to cite: Rahmsdorf, E., Doktor, D., Feilhauer, H., Faude, U., Fichtner, J., and Lange, M.: Large-scale tree species maps for tree species diversity monitoring across scales – challenges and opportunities, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-376, https://doi.org/10.5194/wbf2026-376, 2026.
Imaging spectroscopy has become an essential tool for biodiversity monitoring, enabling the characterization of ecosystem composition, function, and diversity through spectral information, grounded in the spectral variation hypothesis. However, the growing diversity of sensors and platforms introduces both new opportunities and challenges for consistent biodiversity assessment from local to global scales. Differences in spectral resolution, spatial resolution, and signal-to-noise ratio can substantially affect the measured spectral response of the earth’s surface and, consequently, the estimation of spectral diversity. To address these issues, this study conducts a systematic cross-sensor comparison focusing on emerging spaceborne hyperspectral systems (PRISMA and EnMAP), the spaceborne multispectral system (Sentinel-2), and the airborne imaging spectrometer (AVIRIS-4), aiming to evaluate differences in spectral characteristics and diversity estimation for long-term and large-scale biodiversity monitoring. Specifically, a multi-source reflectance dataset was acquired over the Åmose wetland restoration landscape in Denmark in August 2025 and processed with BRDF and geometric co-registration corrections to eliminate undesired view-sun-angle and geometric differences between sensors. Then, spectral intercomparison was performed both for images before and after BRDF correction under unified spatial and spectral resolution by extracting reflectance spectra from several representative land use classes (including nature, intensive agriculture, extensive agriculture, coniferous forest, deciduous forest, water, and built-up) to identify reliable wavelength ranges for subsequent spectral diversity estimation. For the spectral diversity estimation, the high-resolution AVIRIS-4 data (1 m) were resampled to multiple coarser resolutions (e.g. 10, 20, 30, and 60 m). Spectral diversity metrics were computed at each resolution and compared with corresponding spaceborne sensors (PRISMA and EnMAP at 30 m and Sentinel-2 at 10 m and 20 m), revealing how spatial scale and sensor characteristics influence the estimation of spectral diversity. In addition, uncertainty analysis was employed to quantify the variability and assess the reliability of both spectral intercomparison and diversity estimation. In general, this study provides a framework for clarifying the differences in spectral response and spatial resolution of promising sensors for spectral diversity estimation to support consistent biodiversity monitoring. The outcomes contribute to the development of standardized approaches for cross-sensor spectral diversity estimation and uncertainty reduction in next-generation remote sensing of biodiversity.
How to cite: Wu, Z., Normand, S., Hueni, A., Vögtli, M., and Schneider, F.: From Multispectral to Hyperspectral: Cross-Sensor Intercomparison of Spectral Characteristics and Diversity Estimation in the Åmose Wetland Restoration Landscape, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-380, https://doi.org/10.5194/wbf2026-380, 2026.
Global-scale land-use and land-cover (LULC) datasets are essential for predicting biodiversity futures and developing solutions to the related challenges of biodiversity loss, climate change, and food security. Existing harmonised LULC datasets, that is without discontinuities when moving from past to present, have coarse spatial and temporal resolutions that do not match the scale on which environmental processes occur. Additionally, current harmonised LULC datasets do not consider landscape patterns, which are important for processes such as species movement and hydrological dynamics.
We present a downscaled global LULC dataset for five future socioeconomic and climatic scenarios, with 0.01° spatial (approximately 1 km at the equator) and yearly temporal resolutions, that is harmonised with historic LULC to span the period 1960 to 2100. Future LULC projections were generated by downscaling LULC change from a global land system model, LandSyMM, from 0.5° to 0.01° using the LandScaleR algorithm. Prior to downscaling, we calibrated LandScaleR with historic LULC data to ensure that it produced realistic landscape patterns. The spatial variation of historic LULC change was used to inform the spatial patterns of future change during the downscaling process. Calibration of LandScaleR revealed substantial regional variation in past patterns of LULC change. The future LULC projections indicate significant landscape change, and emphasise the importance of incorporating local-scale processes in global LULC projections.
Our harmonised LULC dataset will be beneficial for studying the impacts of LULC change on biodiversity because it was calibrated to ensure that future LULC projections have realistic landscape patterns. The dataset also has high spatial and temporal resolutions that better match the scale of environmental processes compared to existing products. We anticipate that our harmonised dataset will facilitate the integration of LULC with a range of environmental models, including those that model biodiversity, ecosystem services, fire, and hydrology. Overall, this new dataset offers an important tool for informing spatial planning and policy design, and addressing challenges including biodiversity loss and climate change.
How to cite: Woodman, T., Arendarczyk, B., Winkler, K., Henry, R. C., Eigenbrod, F., Burslem, D. F. R. P., Alexander, P., and Travis, J.: High-resolution global land-use maps from 1960 to 2100 for biodiversity modelling, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-487, https://doi.org/10.5194/wbf2026-487, 2026.
Flowering is a foundational ecological process that shapes biodiversity, ecosystem functioning, pollination networks, and agricultural productivity. Yet, despite its centrality, global flowering dynamics remain one of the least observed biological phenomena, historically accessible only through sparse ground observations or localized field studies. Recent advances in flowering spectroscopy—enabled by a new generation of imaging spectrometers—offer an unprecedented opportunity to detect floral signals from the air and space, opening a new frontier for biodiversity monitoring inspired by how pollinators themselves perceive the world. The Spectral-Based Flowering Monitoring System (SFMS) introduces a novel, integrated framework that combines social signals, hyperspectral observations, machine learning, and community-driven validation to map flowering events across global ecosystems, leveraging the emerging capacity of airborne and satellite imaging spectrometers to capture the subtle spectral fingerprints of flowers that were previously undetectable at large scales. SFMS consists of three synergistic components. The Bloom Alert module continuously tracks real-time trends in multilingual social media streams related to flowering, using keyword filtering to locate emergent bloom events reported by the public. These crowdsourced observations guide geolocated targeted analyses and form a continuously expanding archive for downstream validation. Simultaneously, airborne and satellite observations from platforms including AVIRIS, EMIT, PACE, Landsat-8/9, and Sentinel-2A—accessed through NASA's DAAC cloud capabilities—are automatically queried for coverage over both historically documented and newly detected bloom areas. The Spectral Detection & Mapping module operates in two stages. First, a spectral unmixing algorithm decomposes subpixel spectral variability using a residual-based method informed by an extensive floral spectral library derived primarily from AVIRIS-family airborne campaigns. Second, the resulting spectral residuals feed an unsupervised Gaussian Mixture Model that identifies flowering pixels and quantifies their associated uncertainty, enabling spatially explicit flowering extent maps. Finally, the Validation component cross-checks detected blooms with independent ground observations sourced from citizen-science platforms such as iNaturalist, along with very high-resolution satellite imagery from NASA’s Commercial Smallsat Data Acquisition (CSDA) program. By disentangling floral spectral signatures and revealing flowering patterns at landscape to regional scales, SFMS enables new pathways for producing spatial indicators of habitat condition, flowering species distributions, and ecological change driven by climate-related phenology shifts and land-use change.
How to cite: Angel, Y., Kathuria, D., Lang, E., and Shiklomanov, A. N.: Tracking Global Bloom Dynamics from Flower to Orbit: The Spectral-based Flowering Monitoring System, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-544, https://doi.org/10.5194/wbf2026-544, 2026.
The escalating global increase in anthropogenic pressure urgently demands consistent and high-resolution monitoring tools. Protecting nature begins with the ability to accurately track where human influence is intensifying and where natural ecosystems remain intact. Until now, the mapping of human influence has relied on traditional methods combining coarse pressure layers (such as nighttime OpenStreetMap, and land cover) with fixed heuristic rules. These input layers are often imperfect, not temporally consistent, and have known representation biases, particularly in the Global South, severely limiting their utility for conservation policy.
This work proposes the task of Naturalness Mapping from Space. This is an end-to-end geospatial AI framework that learns the complex mapping directly from a single satellite image to quantify the absence of modern human influence.
Our framework first used a feature attribution method adopted from interpretable machine learning domain across multi-modal inputs --including Sentinel-1, Sentinel-2, nighttime lights, and land cover-- to reveal that Sentinel-2 is the most influential modality for the task [1]. Using this insight, we designed a geospatially-aware deep learning model, tailored to EO specifics by conditioning it on geocoordinates and wider spatial context. This results in a high-resolution 10 meters, globally-consistent Naturalness Map, quantifying the degree of human influence.
The final model, deployed at a continental scale, is wrapped by a Distribution Shift Detector [2]. This yields two outputs: the primary Naturalness Map and its accompanying Reliability Map. The Reliability Map shows whether an end-user should trust the model's prediction at a certain area or time, directly quantifying model confidence. This companion product is engineered to increase transparency in AI predictions, ensuring downstream decision-making processes acknowledge model limitations and actively mitigate the risks of misinterpretation stemming from the known biases of initial training data proxies. Our work thus provides a robust, reliable, and trustworthy blueprint for creating large-scale wall-to-wall geospatially-tailored deep learning maps in support of global conservation targets.
[1]B.Ekim,T.T.Stomberg, R.Roscher,M.Schmitt,"MapInWild:A Remote Sensing Dataset to Address the Question of What Makes Nature Wild"in IEEE Geoscience and Remote Sensing Magazine,doi:10.1109/MGRS.2022.3226525
[2]B.Ekim,G.A.Tadesse,C.Robinson,G.Hacheme,M.Schmitt,R.Dodhia,J.M.L.Ferres;Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,2025,pp.2290-2299
How to cite: Ekim, B. and Schmitt, M.: Quantifying Anthropogenic Pressure at the Continental Scale from Space: A Transparency-Aware Geospatial Deep Learning Framework, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-782, https://doi.org/10.5194/wbf2026-782, 2026.
Biodiversity is being lost at an unprecedented rate worldwide. To better address this crisis, efficient methods for monitoring biodiversity are essential. At the same time, Earth-observation satellites provide vast amounts of data that can be leveraged with the help of artificial intelligence. Here, we present a pipeline that uses Sentinel-2 Level-2A imagery and a convolutional neural network (CNN) to predict the potential distribution of individual species.
Our approach ingests all 12 spectral bands from multiple unprocessed images of the same location. The workflow begins with a dataset containing presence–absence records and their coordinates for the species of interest, along with an optional definition of the target prediction area. The pipeline then automatically downloads the required satellite imagery products. For each presence–absence record, we use satellite images from the corresponding year of data collection. To train the CNN model, we extract an 18 x 18-pixel neighbourhood centered on each record location. The pipeline then trains the CNN and generates probability-of-occurrence predictions. Predictions are provided as georeferenced TIFF images, with probabilities computed for each 10 x 10 m pixel.
Two of the major threats to natural ecosystems are the spread of invasive species and the extinction of rare endemics. Therefore, we demonstrate the utility of our approach using two example species: an invasive mosquito and an endemic plant. The Asian tiger mosquito (Aedes albopictus), a recently established species in Europe, can act as a vector for several pathogens (e.g., yellow fever). Training data for this species were obtained from the citizen-science programme szunyogmonitor.hu. Our second case study focuses on Seseli leucospermum (“magyar gurgolya”), a strictly protected plant species endemic to Hungary, for which records were obtained from the biodiversity databases of Hungarian national parks. We successfully trained our model for both species, achieving validation accuracies above 80% and now calculate the predictions. Future work will include field validation of the predictions.
This study was supported by the Hungarian Research, Development and Innovation Office (grant K138503).
How to cite: Barta, Z., Gáspár, Á., Bán, M., Garamszegi, L. Z., Bérces, S., Szabó, S., and Barta, A.: Deep learning meets Sentinel imagery to predict species distributions., World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-223, https://doi.org/10.5194/wbf2026-223, 2026.
Monitoring how birds, bats and insects move through the airspace is essential for biodiversity assessment and for understanding how different taxa use this aerial domain. Radar has long enabled continuous observations of animal movements, but the lack of species- or trait-specific information has limited its ecological value. Recent advances in FMCW radar provide detailed micro-Doppler spectrograms that capture wingbeat dynamics with high temporal resolution, creating new opportunities for taxonomic differentiation. Yet these signatures remain difficult to interpret, because they are not directly linked to known biological traits or species-specific reference data.
We present a physics-based simulation and inference framework that links articulated animal models directly to radar observables. The simulator describes detailed wing and body kinematics, including stroke geometry, flapping dynamics and morphology, and propagates these motions through radar backscatter model to generate realistic micro-Doppler signatures. These synthetic radar returns provide the basis for training an inference module that learns to recover the underlying kinematic parameters from field spectrograms. The recovered parameters, when passed through the forward simulator, reconstruct radar signatures that closely match measured data, showing that physically grounded supervision enables inverse kinematics without manual labels and produces biologically interpretable descriptors of flight.
We apply the framework to a vertical FMCW radar dataset capturing birds, bats and insects moving through a fixed airspace column. When run on these unlabeled field recordings, the model retrieves trait-based descriptors that expose systematic differences in flight behaviour across taxa and allow tracks to be organized into biologically meaningful groups. This demonstration shows that physically grounded interpretation can reveal structure that is otherwise inaccessible in raw micro-Doppler data, offering a practical basis for continuous, multi-taxon monitoring of aerial biodiversity. By linking radar physics to biological traits in a scalable, non-invasive way, the approach bridges the gap between radar physics and ecology for monitoring avian biodiversity and expanding the biological interpretability of radar observations.
How to cite: Liechti, F., Nüesch, J., Aguettaz, H., and Kleger, D.: From Radar Blips to Species-Level Aerial Monitoring: Physics-Based Simulation Enables Biological Insight from Micro-Doppler Data, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-478, https://doi.org/10.5194/wbf2026-478, 2026.
Energy–biodiversity hypotheses aim to explain the mechanistic relationships between energy availability in the environment and the processes that generate and maintain species richness. These hypotheses propose that the amount, variability, and minimum levels of available energy can influence speciation, extinction, and the overall capacity of ecosystems to support diverse biological communities. The Available Energy (Cumulative) Hypothesis suggests that ecosystems receiving greater annual cumulative energy inputs can support more species because higher energy availability increases resource production. This hypothesis is supported by both in situ field observations and controlled in vivo studies. The Environmental Stability (Variation) Hypothesis proposes that lower intra-annual variability in energy promotes species richness by providing predictability and reducing physiological stress, which enables more species to coexist. In contrast, the Environmental Stress (Minimum) Hypothesis emphasizes the role of minimum energy thresholds, suggesting that regions maintaining higher minimum levels of energy throughout the year can support more species by exceeding the baseline physiological requirements necessary for survival and reproduction. These three hypotheses (cumulative, variation, and minimum energy) collectively explain between one-half and two-thirds of the geographic distributions of amphibians, mammals, and birds in terrestrial systems, demonstrating their broad explanatory power. Importantly, the mechanisms underlying these hypotheses are not inherently restricted to land-based ecosystems. In this study, we extend these ideas to the marine environment to assess whether similar energy–biodiversity relationships emerge in the ocean. Using satellite-derived observations, we developed two radiative energy indices (photosynthetically active radiation and sea surface temperature) and three metabolically based indices (primary production and benthic particulate organic carbon flux). We paired these indices with species richness data for marine fish, mammals, reptiles, lobsters, abalone, conus species, corals, and seagrasses. Our results show that radiative energy indices explained up to 63% of the variation in species richness for certain taxa, whereas metabolic indices were generally less predictive. As in terrestrial ecosystems, cumulative energy was most important offshore, while energy variation more strongly influenced coastal biodiversity, likely reflecting the dynamic and productive nature of coastal habitats. Collectively, our findings demonstrate that energy availability is closely linked to global patterns of marine species richness.
How to cite: Oliver, M., Zmina, S., Reygondeau, G., and Radeloff, V.: Testing Energy Biodiveristy Theories on Marine Species, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-571, https://doi.org/10.5194/wbf2026-571, 2026.
Collecting in-situ environmental data and biological samples in polar regions remains extremely challenging. Manually obtaining samples near large icebergs and water-facing glaciers is especially difficult, as the melting ice poses physical dangers for researchers. Icebergs can tip without warning and glaciers calving can create unpredictable landslides and tsunamis. Due to these risks, data is scarce, and little is known about the microbiology at the glacier-ocean interface. Yet these environments are ecologically dynamic and face unprecedented environmental change; iceberg-triggered upwellings, glacier runoffs, and drifting icebergs mix the water column, alter stratification and nutrient availability, and thereby reshape microbial community composition, primary productivity and carbon cycling. Similarly, glacier runoff releases previously trapped microbes.
To help fill these observational gaps, we present MEDUSA, an aerial-aquatic robot capable of flying to far locations and performing underwater sensing. The newest variant of the MEDUSA robotic family, introduced here, combines flight with underwater sensing and sampling, enabling CTD profiling, targeted water retrieval, and filter sampling. This robot will enable new methods of obtaining water samples in extreme environments and could have a sweeping impact as to the amount of data available in Polar research. It can also support the interpretation of remote sensing measurements and strengthen multi-sensor biodiversity monitoring frameworks. Beyond polar regions, MEDUSA will be deployed in Swiss lakes and Mediterranean environments, supporting scalable biodiversity monitoring across diverse aquatic ecosystems.
Recent field trials in southern Greenland, conducted from the Forel Research Platform, demonstrate MEDUSA’s ability to operate safely in iceberg-dominated fjords and obtain samples. The Forel is a sailing research boat, with an onboard clean-room, chemistry lab, and workshop, designed for environmental research. The boat approached glaciers and large icebergs, with MEDUSA closing the distance and sampling the water without posing a human risk. Here, we introduce our sampling methodology, as well as the hardware results of our field trials. We discuss the challenges and insights from this deployment, as well as the future directions and advancements for this robotic platform.
How to cite: Holdcroft, K., Romanello, L., Orr, L., Wolfart, G., Taureg, A., Paoli, L., Louchard, D., and Kovac, M.: Aerial–Aquatic robotics for safe and scalable biodiversity monitoring in polar environments , World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-607, https://doi.org/10.5194/wbf2026-607, 2026.
The accelerating evolution of remote sensing technologies offers unprecedented opportunities to assess biodiversity across ecosystem structure, composition, and function. Within this landscape, our BlueMAP (Blue Carbon Market Acceleration Potential by EnMAP) project demonstrates how advanced multi-sensor Earth Observation combining hyperspectral EnMAP data with multispectral Sentinel-2 and PlanetScope imagery can substantially enhance biodiversity monitoring in dynamic coastal ecosystems such as mangroves and seagrass meadows. These habitats are globally significant for biodiversity and carbon sequestration, yet remain challenging to monitor due to their fine-scale spatial heterogeneity, water-column effects, and rapid ecological dynamics.
BlueMAP develops and tests new remote-sensing workflows that translate multi-sensor EO-derived information into biologically meaningful metrics. Through hyperspectral unmixing, data fusion, and machine-learning methods, the project improves discrimination of benthic and wetland habitat classes, enabling enhanced mapping of ecosystem extent, structural traits (e.g., canopy density, seagrass cover), and indicators of ecosystem condition. The integration of EnMAP’s hyperspectral richness with the temporal frequency of Sentinel-2 and PlanetScope enables identification of subtle ecological patterns such as degradation signals in mangroves or annual density shifts in seagrass beds supporting more accurate derivation of biodiversity-relevant variables including ecosystem structure, health, and functional proxies.
A key innovation of BlueMAP is its direct alignment with the needs of emerging blue-carbon markets, providing robust, transparent, and scalable ecosystem monitoring products that reduce uncertainty in carbon credit generation and verification. By integrating hyperspectral EnMAP data with high-resolution multisensor imagery, BlueMAP delivers consistent measurements of ecosystem extent, condition, and above-ground biomass, critical inputs for Tier-2 carbon stock estimation and MRV workflows. This EO-based approach improves the ability to separate intact from degraded habitats, monitor ecological change, and quantify biomass dynamics. In doing so, it tackles key challenges in biodiversity remote sensing, such as ensuring consistency across sensors, reducing uncertainties, and turning satellite derived information into ecosystem metrics that are ready for policy use and blue-carbon markets.
BlueMAP illustrates how next-generation EO missions and multi-sensor integration can strengthen biodiversity monitoring frameworks that underpin credible blue-carbon markets and emerging nature-based climate finance, while also aligning with broader policy agendas such as the Kunming–Montreal Global Biodiversity Framework and the SDGs.
How to cite: Ressl, R., Hartmann, K., Heege, T., Osterkamp, N., and Bauer, R.: Advancing Coastal Biodiversity Monitoring and Blue-Carbon Metrics through Hyperspectral-Multisensor Earth Observation, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-726, https://doi.org/10.5194/wbf2026-726, 2026.
Rapid changes in ocean climate, circulation, and ecosystem structure are driving large-scale shifts in marine species distributions—challenging biodiversity conservation and the sustainability of fisheries worldwide. Anticipating these shifts is central to climate-ready ecosystem management, yet many existing species distribution modeling (SDM) approaches struggle to capture non-linear ecological responses, multi-scale environmental drivers, and emergent behaviors that arise under rapid climate change. To support proactive, climate-ready management, we developed an interpretable AI framework that integrates satellite remote sensing, high-resolution oceanographic reanalyses, and in situ data streams from a range of species spanning forage fishes to top predators. We use this framework to forecast the dynamic spatial and temporal (re)distribution of marine species in the Northwest Atlantic Ocean under variable oceanographic conditions. By combining multi-sensor observations with explainable machine learning methods, our approach provides both high predictive accuracy and transparent, ecologically grounded insights into the drivers of species movement. We compare the performance and usability of interpretable AI models with traditional statistical approaches. Whereas conventional SDMs typically assume smooth, stationary relationships between species occurrence and environmental predictors, interpretable AI models more flexibly capture threshold effects, interacting climate drivers, and regionally varying habitat relationships. We demonstrate how this system identifies climate-resilient habitats, reveals emerging hotspots of overlap between fisheries and vulnerable species, and dynamically informs spatial management measures such as time–area closures, protected area design, and climate sentinel site monitoring. Working directly with management agencies and protected area networks, we co-develop actionable metrics of ecosystem climate vulnerability and translate model outputs into operational decision support tools. This work illustrates a scalable pathway for uniting remote sensing, AI, and in situ observing networks to monitor biodiversity change and its drivers. By aligning technological advances with ecological understanding and management needs, our approach contributes directly to adaptive marine spatial planning and supports implementation of related domestic and global policy targets.
How to cite: Braun, C., Farchadi, N., McDonnell, L., Mcdermott, J., Milles, H., and Lewison, R.: Using interpretable AI to integrate remote sensing and in situ data for forecasting marine species and climate-ready marine spatial planning, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-360, https://doi.org/10.5194/wbf2026-360, 2026.
The NASA Global Ecosystem Dynamics Investigation (GEDI) is a spaceborne light detection and ranging system (LiDAR) focused on characterizing the three-dimensional structure of Earth’s temperate and tropical forests. A primary objective of GEDI is to assess the linkages between forest structure, habitat quality, and biodiversity at large spatial scales. Here, we introduce a project that used data fusion frameworks to create continuous GEDI-fusion forest structure metrics at 30 m across a large swath of the western United States. We leveraged these products in three separate studies to examine GEDI’s ability to assess habitat quality, biodiversity patterns, and species interactions. The first study used capture-mark-recapture data of two small mammal species to estimate parameters critical to habitat quality, density and survival. We found strong support that densities were associated with GEDI-fusion forest structures but weak support that survival was associated with GEDI-fusion metrics. We then used these findings to create spatially explicit density maps to aid management and conservation policies. The second study evaluated how GEDI-fusion metrics influence bat occupancy and diversity across broad environmental gradients and identified whether species–habitat relationships were stationary or nonstationary (i.e., fixed vs vary spatially). Using data from the North American Bat Monitoring Program and multispecies occupancy models we found GEDI-derived forest structures were influential, but nonstationary, drivers of individual species occupancy processes that scaleup to shape bat diversity patterns across the region. We demonstrate that incorporating GEDI and flexible spatial models can better support biodiversity assessments across broad, ecologically heterogeneous landscapes. The final study used GEDI-fusion products to predict the occupancy and species interactions of a carnivore guild through Bayesian occupancy models and structural equation modeling. GEDI-fusion forest structures were critical drivers of occupancy and mediated interactions between competitors and predators of two protected species. Collectively, this project represents novel applications of GEDI data and we conclude that forest structure characterized via GEDI data fusions can be used to assess demography, diversity, and interactions of terrestrial mammals.
How to cite: Barry, B., Holbrook, J., Vogeler, J., and Vierling, K.: Wildlife Applications of GEDI, Spaceborne LiDAR, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-857, https://doi.org/10.5194/wbf2026-857, 2026.
In 2019, an alarming study of bird populations revealed that 3 billion birds have been lost since the 1970s. Bird populations are predicted to decrease further as a result of climate change and land conversion. Additionally, the usefulness of bird diversity as an indicator of overall biodiversity and ecosystem health has been well researched, as has the connection between forest structure and avian species. However, on-the-ground studies of forest structural diversity are often limited by time and financial resources. Remote sensing of forest structure, such as discrete return and aerial lidar, demonstrably improve species model performance in regional studies to predict bird species occurrence. In addition, the ability of spaceborne lidar to improve biodiversity predictions is still being explored and offers a publicly available method for measuring forest structure across large regions. Here, we will use the Global Ecosystem Dynamics Investigation (GEDI) lidar instrument to improve existing species habitat suitability models and predict biodiversity hotspots in a National Park in Africa. GEDI is a space-borne lidar instrument aboard the International Space Station that is capable of measuring the height and complexity of vegetation. In addition to sociocultural and ecological data, we will compile GEDI derived metrics of forest structure, including canopy height (RH98), foliage height diversity (FHD), plant area index (PAI), waveform structural complexity index (WSCI) to apply structural variables to an ensemble model of species distribution. Using species occupancy gathered from in-situ point count data from 2022-2023, along with the SSDM package in R, we will create stacked species distribution models and species richness predictions. Models with and without GEDI data will be compared to understand the impact of GEDI metrics on model accuracy. The results of this work will inform park management and bolster efforts to conserve species biodiversity in the park using remote sensing tools. Results of this work will also inform biodiversity hotspot mappings across larger regions of Africa where in situ data is sparse.
How to cite: Barenblitt, A., Stovall, A., and Duncanson, L.: Spaceborne Lidar is for the Birds: Applications of GEDI lidar to improve species distribution and hotspot mapping in a National Park in Africa, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-900, https://doi.org/10.5194/wbf2026-900, 2026.
Accurately quantifying biodiversity at scale is critical for conservation, particularly as new conservation initiatives require robust metrics to assess and monitor ecological outcomes. Advances in remote sensing, analytical methods, and expanding in situ observation networks now provide unprecedented capacity to model biodiversity at fine spatial resolution across large regions. Historically, large-scale biodiversity assessments relied on coarse proxies such as climate, land cover extent, or expert knowledge, which can obscure species-level responses. There is an urgent need for approaches that link biodiversity outcomes to species explicit habitat estimates and reveal environmental drivers of biodiversity patterns. We present a high-throughput species distribution modeling workflow that provides a species explicit approach to quantifying biodiversity as a function of fine-scale habitat suitability. Using Google Earth Engine and a machine learning framework, we evaluated how high resolution, multi-scale environmental covariates, primarily derived from remote sensing and representing climate, land cover, disturbance, and landscape configuration, influence species distributions. We then stacked species models to generate regional biodiversity layers that reflect species-level habitat suitability rather than coarse habitat surrogates. We applied this approach to 191 bird and mammal species across the North American Great Plains using occurrence data from eBird and GBIF. For each species, we implemented a covariate selection process to identify the optimal spatial scale for each predictor and built final models using only the best-scale covariates. Model performance was consistently high, producing detailed spatial predictions within expected species’ ranges. The models reveal how species respond to environmental drivers across scales, providing nuanced insights into biodiversity patterns in heterogeneous landscapes. While we are using this workflow to assess and monitor biodiversity outcomes from rangeland restoration on private lands, it is readily extendable. It can support hindcasting to determine drivers of biodiversity change, multispecies connectivity modeling, or integration with trait-based indices for functional diversity. The derived species richness layers provide a scientifically rigorous, spatially explicit measure of biodiversity that can guide conservation prioritization and management, scenario planning, and biodiversity finance strategies. By moving beyond coarse proxies, this approach offers a scalable, species-centered method for assessing and managing biodiversity across regional and global scales.
How to cite: Hightower, J., Lacey, M., Standen, M., and Suraci, J.: A high-throughput, species-explicit approach to quantifying biodiversity at scale for assessment and monitoring, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-840, https://doi.org/10.5194/wbf2026-840, 2026.
Remote sensing plays an increasingly central role in biodiversity monitoring, yet the robustness of remotely derived biodiversity products depends on our ability to quantify, understand, and propagate the uncertainties embedded across the sensing and processing chain. The uncertainties arise at multiple stages in the remote sensing workflow, ranging from sensor noise and atmospheric interference to processing algorithms, model assumptions, and instrument calibration and validation. Despite this, biodiversity assessments rarely incorporate uncertainty information, limiting the reliability, comparability, and interpretability of biodiversity measurements and change assessment. Understanding how these uncertainties affect biodiversity measurements and indicators has potential implications for our understanding of biodiversity processes and the use of this information in decision-making.
As a growing number of remote sensing missions are about to deliver pixel-level uncertainty estimates, the field faces an important challenge: how to move towards integrating uncertainty directly into biodiversity assessments, conservation and restoration planning, and policy-relevant indicators.
In this contribution, we propose a roadmap to operationalize the integration of uncertainty in biodiversity product generation. We outline key stages where uncertainty can be quantified and propagated, highlight conceptual considerations for different types of biodiversity products, and demonstrate how uncertainty-aware workflows can support more transparent and robust assessments.
We test this approach using EMIT (Earth Surface Mineral Dust Source Investigation) imaging spectroscopy data and its accompanying uncertainty data in temperate forest ecosystems. Leveraging EMIT’s per-band standard deviation layers, we propagate surface reflectance uncertainty into plant traits and compare these uncertainty-aware EMIT-derived plant traits with in situ leaf spectroscopy and laboratory plant traits. This application aims to start a dialogue on how to move towards uncertainty-aware biodiversity products and indicators that ensure measurement robustness and identify where caution is needed.
Our results underscore that explicit inclusion of uncertainty is fundamental for biodiversity monitoring. By embedding uncertainty into both product generation and interpretation, we enhance transparency, strengthen ecological inference and knowledge, and support informed decision making.
How to cite: Koch, T. L., Rossi, C., Hueni, A., Karaman, K., and Santos, M. J.: Integrating uncertainty into remotely sensed biodiversity products , World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-771, https://doi.org/10.5194/wbf2026-771, 2026.
Quantifying plant diversity from space remains challenging due to the traditionally coarse spatial and spectral resolution of remotely sensed data relative to the inherent vegetation canopy variability. Traditional spectral diversity metrics assume homogeneous pixel composition and fail to account for sub-pixel heterogeneity, limiting their effectiveness in complex ecosystems, including the tropics. Spectral unmixing offers a promising solution by decomposing mixed pixels into pure spectral signatures (endmembers) and their corresponding abundances, enabling species-level identification and diversity assessment. Previous studies have demonstrated the potential of endmember diversity (EndDiv) as a proxy for plant diversity in grassland using simulated spaceborne imaging spectrometers applied to 1-m airborne data. However, the approach's effectiveness in dimensionally complex tropical forests remains unexplored. The Airborne Validation Unified Experiment: Land to Ocean (AVUELO) campaign provides an opportunity to address this knowledge gap through its comprehensive dataset of 1-m and 3-m airborne imaging spectroscopy data collected with AVIRIS-3 and across diverse tropical forest types. This study applies spectral unmixing to simulated spaceborne imaging spectrometers using multi-resolution airborne spectroscopy data from Barro Colorado Island (BCI). This airborne data is convolved to match existing and planned spaceborne imaging spectrometer systems to later calculate endmember diversity across BCI. We compare derived endmember diversity metrics with field-based biodiversity indices, including spectral, functional, taxonomic, and phylogenetic diversity, to evaluate the efficacy of spaceborne imaging spectroscopy missions by: (1) assessing the effectiveness of endmember diversity for quantifying plant diversity in tropical forests, (2) determining optimal spatial and spectral resolution requirements for biodiversity monitoring, (3) establish relationships between EndDiv and multiple dimensions of biodiversity in this Neotropical forest, and to finally (4) compare it with other remotely sensed biodiversity metrics. The results will inform the design and application of current and future spaceborne imaging spectroscopy missions for global biodiversity monitoring, particularly in tropical ecosystems where biodiversity assessment is challenging.
How to cite: Quinteros Casaverde, N., Angel Lopez, Y., Baresch Aristizabal, A., Czyz, E., Wright, J., Muller-Landau, H., Townsend, P., Brodrik, P., Podest, E., Schimel, D., and Serbin, S. and the AVUELO Support Team: AVUELO Endmember Diversity, Its Relationship with the Dimensions of Biodiversity, and Implications for Current and Future Imaging Spectroscopy from Space, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-793, https://doi.org/10.5194/wbf2026-793, 2026.
Satellite remote sensing capabilities have advanced rapidly in recent years, with multispectral optical imagery, LiDAR, synthetic aperture radar (SAR), and thermal observations now providing unprecedented potential to assess biodiversity-relevant habitat characteristics at national to global scales. These measurements directly support Essential Biodiversity Variables (EBVs)—especially Ecosystem Vertical Structure, Phenology, and Primary Productivity—that underpin analyses of species–habitat relationships, ecological processes, and responses to climate and land-use change. Yet despite numerous successful demonstrations over limited extents, operational biodiversity data products remain scarce.
A major constraint is the lack of community-agreed product specifications, including spatial resolution, temporal frequency, vertical accuracy, uncertainty thresholds, and standardized formats required for broad adoption by biodiversity scientists, wildlife managers, and policy stakeholders. This impedes the development of globally consistent Biodiversity Indicators aligned with reporting needs under the Kunming–Montreal Global Biodiversity Framework (GBF).
To help close this gap, NASA’s Jet Propulsion Laboratory (JPL) has initiated a structured series of engagement activities to define priority satellite products that would deliver the greatest value for biodiversity conservation. Here, we present findings from the Remote Sensing for Animal Movement (RSAM) study, synthesizing perspectives from more than 50 experts in movement ecology and satellite remote sensing to:
-
identify habitat characteristics most critical for current animal movement studies and define required spatial, temporal, and vertical resolutions;
-
develop an ecologically meaningful typology of habitat attributes relevant to movement processes;
-
assess whether current satellite systems meet these observational needs and where measurement bottlenecks persist; and
-
highlight opportunities for next-generation capabilities to fill gaps in spatial detail, measurement frequency, and environmental representation.
Across terrestrial, freshwater, and marine systems, participants consistently ranked Ecosystem Structure and Ecosystem Function as the highest-priority observation needs, with Composition, Condition, and Use following in importance.
Sustained engagement with additional biodiversity end-user communities will extend these findings into a coordinated portfolio of analysis-ready, routinely produced satellite biodiversity products, enabling actionable conservation decisions and transparent global biodiversity assessment.
How to cite: Ferraz, A.: Satellite Data Products: Community Priorities for Biodiversity Science and Conservation, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-928, https://doi.org/10.5194/wbf2026-928, 2026.
Launched in February 2024 and serving data to the public as of April 2024, the NASA Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) satellite provides a novel set of daily, hyperspectral and polarimetric Earth observation (EO) capabilities that are unmatched by any other satellite platform, public or private. PACE's primary sensor, the Ocean Color Instrument (OCI), is a hyperspectral (5 nm resolution) imaging spectrometer that measures properties of light from the ultraviolet (UV, 340 nm) through visible and near infrared (NIR, 895 nm) portions of the electromagnetic spectrum in 1.25 to 2.5 nm steps, plus 7 shortwave infrared (SWIR) bands. Observations from OCI are moderate spatial resolution (1.2 kmx 1.2 km), enabling a relatively frequent revisit time of 1-2 days. OCI is complemented by two cubesat-sized polarimeters, the hyperspectral Spectrometer for Planetary Exploration (SPEXone) and the Hyper-Angular Rainbow Polarimeter (HARP2). PACE is comprehensive Earth System mission, benefiting society by expanding our foundational knowledge of Earth and enabling novel, space-based science and applications tools.
Continuous, synoptic imaging spectroscopy of Earth from PACE provides an opportunity for next-generation biodiversity assessment and environmental monitoring applications tools. For example, photosynthetic phytoplankton (microscopic algae) and terrestrial plants package the Sun’s energy for higher trophic levels, ultimately fueling most aquatic and terrestrial life on Earth. With PACE we can utilize hyperspectral-enabled metrics of their diversity, growth, and photophysiology to monitor how algal and plant type, distribution, and health varies over time. PACE can additionally support applications areas relevant to changes in biodiversity and anthropogenic impacts such as water and land resource management, climate-biodiversity connections, air quality and public health, and more. In this presentation we will provide an overview of the PACE mission and provide real-world examples of PACE data in action. Translation of PACE’s novel observations to actionable, trusted new tools is a long term, continually developing process and supported by our PACE Applications Program. We will provide resources on how you can be involved in PACE Applications activities through our PACE Community of Practice, Early Adopters Program, and information-sharing and co-production workshops and focus sessions that we plan throughout the year.
How to cite: McKibben, S. M.: NASA’s hyperspectral PACE mission: New potential for monitoring biodiversity from space, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-922, https://doi.org/10.5194/wbf2026-922, 2026.
Mountainous regions are characterised by unique biodiversity and provide essential ecosystem services. However, they are also particularly vulnerable to global change. In the Mediterranean mountains, climate change has a significant impact on plant productivity and phenology, through an increase in extreme drought events and changes in snowfall and temperature patterns. Additionally, the progressive abandonment of pastoralism is leading to profound changes in open and semi-open habitats (subalpine grasslands, heathlands and forest ecotones) through the densification and recolonisation of woody species (ericaceous shrubs and conifers in particular). There is an urgent need to develop a more profound understanding of these complex vegetation dynamics in order to inform the development of coherent guidelines for stakeholders in mountainous regions, including livestock farmers and nature reserve managers, to help them adapt their practices to preserve both foraging resources and rich biodiversity. In order to monitor past and current vegetation dynamics in open and semi-open habitats, we propose applying a method of classifying ecological trajectories into nine highly interpretable categories. This method involves adjusting a second-degree polynomial function and provides insight into the direction and acceleration of the studied trajectories (Rigal et al., 2020). To describe changes in vegetation functioning and composition, we calculated the Dynamic Habitat Index (DHI) using Landsat and Sentinel-2 time series data. Finally, we aimed to identify the impact of potential climate-related drivers in different contexts of agro-pastoral management. Our approach was applied to the alpine and subalpine landscapes of the Mediterranean Pyrenees in southern France, with the results highlighting the existence of differential trajectories depending on the habitat considered, and potentially the pastoral management practices in Catalan nature reserves with high biodiversity values. Our results encourage us to move beyond traditional remote sensing approaches based on calculating greening/browning trends to monitor vegetation dynamics patterns, providing a more detailed framework for analysing the complex changes affecting mountain ecosystems.
How to cite: Defossez, A., Thierion, V., Alleaume, S., Berchoux, T., and Luque, S.: Monitoring complex vegetation trajectories in open and semi-open habitats of the Mediterranean Pyrenees, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-120, https://doi.org/10.5194/wbf2026-120, 2026.
Recent advances in large vision–language models (VLMs) offer new opportunities to extract structured, biologically relevant information from unstructured, image-like data, including earth observation satellite and ocean observation marine remote sensing modalities. We present a lightweight pipeline that uses few-shot, instruction-aligned VLMs to convert visual inputs into concise, schema-based JSON records capable of capturing specific, but flexible scene content such as habitat characteristics, disturbance events, species labels, data-quality attributes, Essential Biodiversity Variables (EBV) or any user-defined ecological indicators. This system enables rapid generation of consistent annotations and structured data extraction across massive archives without task-specific model training, complex engineering, or human labels. We demonstrate the usefulness of this approach through practical applications to automated quality-control tagging, low-level land-use and habitat-type classification, species classification, and numerical feature estimation from 2-D images generated by satellite platforms as well as complementary sensors such as shipboard acoustics (EK60) and S-band radar. We optimize performance generally and on a task-specific basis through the incorporation of human-like spatial reasoning via grid-referenced subregion analysis and prompt-optimization frameworks such as DSPy for declarative prompt programming and self-improvement. By producing interpretable, reproducible and harmonized annotations at scale, our approach substantially reduces the manual screening effort required to curate multi-sensor datasets, prioritizes scenes for higher-fidelity processing and supports sophisticated cross-platform analysis aligned with biodiversity applications. VLM technologies are rapidly reshaping environmental data management, and our results provide an early, practical demonstration of how VLM-based visual interpretation can enhance the flexibility, scalability, and interoperability of remote-sensing pipelines for biodiversity monitoring. Moreover, these capabilities directly support key reporting needs under the Kunming–Montreal Global Biodiversity Framework, particularly Targets 1, 2, 4, and 19 that require scalable information on ecosystem extent, condition, disturbance, and data accessibility. They also contribute to SDG indicators (e.g., 15.1.1, 15.3.1, 14.2.1) by enabling rapid, harmonized extraction of habitat, land-use, and marine-ecosystem attributes from multi-sensor Earth observation archives.
How to cite: Culhane, E. and Barnett, E.: Scalable AI‑assisted annotation of remote sensing imagery, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-550, https://doi.org/10.5194/wbf2026-550, 2026.
The Biodiversity Survey of the Cape (BioSCape) is NASA’s first biodiversity-focused airborne and field campaign, designed to advance understanding of ecosystem structure, function, and composition, and how these are changing over time and space. The regional focus is the Greater Cape Floristic Region of South Africa, a biodiversity hotspot. This international collaboration involving ~160 US and South African scientists across 19 PI-led projects produced an unprecedented, open-access dataset spanning >50,000 km², integrating imaging spectroscopy sampling from 0.35 μm to 12 μm (AVIRIS-NG, PRISM, HyTES), LiDAR (LVIS, ELMAP-V), and extensive ground data from ~2,600 field samples. Data are publicly archived through NASA Earthdata and accessible via a cloud computing environment, fostering open science and reproducibility. In this presentation, we will summarize the results to date across the BioSCape project, highlighting key scientific outcomes from both terrestrial and aquatic landscapes. We will also discuss some of the broader implications of the project, including laying the groundwork for achieving ethical and equitable benefit-sharing via compliance with the Nagoya Protocol. A core principle of BioSCape is "use-inspired science," which involves co-developing research questions with local partners to directly address pressing conservation and resource management challenges. This ensures that our scientific findings translate into actionable data products for decision-makers in South Africa and globally. For example, BioSCape's data and products are already informing conservation and resource management efforts, including kelp range maps for ecosystem classification and woody invasive alien plant maps that feed into national ecosystem condition assessments. The project's contributions have been recognized in the 2024 State of Environment Outlook Report for the Western Cape Province of South Africa. This work serves as a demonstration for other regionally-focused biodiversity campaigns, and is directly advancing remote sensing of biodiversity methods for the next generation of satellite-based technologies to enable improved biodiversity measurements globally. It was also the subject of a short documentary film, “The Spectrum of Life,” available at bioscape.io/film.
How to cite: Wilson, A., Hestir, E., Slingsby, J., and Cardoso, A.: Remote Sensing for Biodiversity: Insights from the BioSCape Campaign, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-850, https://doi.org/10.5194/wbf2026-850, 2026.
Human-driven forest cover changes impact carbon sequestration, climate, biodiversity, and various ecosystem services, being at the same time spatially diverse worldwide. Therefore, there is a need to monitor them at multiple scales. A wide range of globally available land cover datasets and products now enables detailed and comprehensive analysis of forest cover dynamics. For example, declassified spy-satellite imagery provides high-resolution insights into forest cover more than 50 years into the past, while contemporary remote-sensing products offer modern, high-resolution forest maps with varying level of thematic detail. In this study, we propose an optimized method for analysing forest cover change, by combining these two opportunities – historical satellite imagery and recent remote sensing thematic products. As a starting point, we tested different strategies to automatically analyse high-resolution CORONA imagery (1.8-7.5 meter spatial resolution) for 1960-1974 by using object based image analysis (OBIA). For the recent period, we evaluated the accuracy of 3 different contemporary global land cover products including: Google Dynamic World (GDW), ESA World Cover map (WC) and Esri Land Cover (ELC). Our analysis was conducted in test areas in Poland, where the recent forest cover increase was substantial. We found that the optimal segmentation approaches and classification strategies offer high-quality CORONA-based forest mask (F1 score = 0.95). For the contemporary forest cover, the highest accuracy was achieved by combining three tested masks rather than by relaying on a single product. Our analysis shows that despite a land cover data deluge observed in the recent years there is a need for critical approach while analysing forest cover change over time.
Acknowledgements:
This research was funded in whole or in part by the National Science Centre, Poland (UMO-2024/53/N/ST10/02518). For the purpose of Open Access, the author has applied a CC-BY public copyright license to any Author Accepted Manuscript (AAM) version arising from this submission.
How to cite: Shahbandeh, M., Kaim, D., and Kozak, J.: Forest cover change – local scale analysis based on globally available datasets, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-531, https://doi.org/10.5194/wbf2026-531, 2026.
Data from satellite remote sensing offer opportunities to estimate functional diversity from trait-related spectral indices in forest ecosystems at landscape scales and with a very short revisiting time. Until now, most studies on remotely sensed functional diversity of vegetated areas have relied on single-date imagery, typically during peak greenness, and therefore neglected seasonal variability. Here, we aim to understand the potential of multi-year dense time series from Sentinel-2 for large-scale biodiversity monitoring.
We examined seven-day Sentinel-2 composites spanning five years (from 2017 to 2021) of about 250 km² of temperate mixed forests in northeastern Switzerland. We quantified temporal patterns during the growing season in three spectral indices (CIre, CCI, and NDMI) linked to physiological canopy traits (canopy chlorophyll content, carotenoid/chlorophyll ratio and canopy water content) and corresponding diversity metrics (richness and divergence) throughout the entire forested area and among different forest types.
Not only the spectral indices but also the resulting diversity metrics showed pronounced seasonal and interannual variation, indicating environmental sensitivity. The diversity estimates often showed deviations from their estimation during peak-greenness conditions, showing that the timing of the measurement has a crucial influence on the resulting diversity maps. We further found that needle-dominated stands exhibited higher overall richness and divergence than broadleaf stands, and divergence showed comparatively stable behavior across years and communities.
Our results show that observations from dense time series are essential for approaches using remote sensing data in biodiversity monitoring and underscore the need for new methods that explicitly account for the temporal dimension of satellite data. The provided approach can complement field-based methods, but new field-based datasets on biodiversity should consider the timing of measurements to complement the temporal aspect of satellite data. Overall, our work contributes to enhancing the capacity of remotely sensed dense time series from Sentinel-2 for long-term biodiversity monitoring and ecosystem resilience assessment under changing environmental conditions.
How to cite: Helfenstein, I., Koch, T., Schuman, M., and Morsdorf, F.: Temporal Dynamics of Remotely-Sensed Functional Diversity in Temperate Forests from Sentinel-2 Time Series, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-735, https://doi.org/10.5194/wbf2026-735, 2026.
Understanding long-term changes in agricultural landscape structure is essential for assessing biodiversity patterns, ecosystem condition and the impacts of land-use policy. Landscape metrics such as field size, edge length and spatial connectedness serve as important biodiversity-relevant indicators in cropland mosaics, yet their derivation over large areas and long time periods remains challenging as it requires consistent high-resolution remote sensing data and a robust method for the accurate delineation of field patches.
Here we exploit the strengths of the Segment Anything Model (SAM) for the automatic segmentation of farmland parcels using two distinct sources of remote sensing imagery: historic panchromatic CORONA data from 1965 and modern digital orthophotos from 2021/2022. Our study focuses on the German state of Saxony, which experienced severe transitions in agricultural landscape structure throughout the 1960s due to substantial policy changes, such as collectivization, with major consequences for field size and associated farmland biodiversity. We selected several study regions across the state for which scenes from both data sources were pre-processed, harmonized and masked. We then employed a two-step SAM algorithm combined with customized post-processing steps and evaluated segmentation performance using the Intersection over Union (IoU) between predicted patches and user-derived validation patches. Input parameters such as compactness and minimum size were tuned to favor “field-shaped” segments. Across most test regions, the method achieved high accuracy with median IoU values of around 0.7 for historic and over 0.9 for modern images, with some limitations in mountainous areas and those with low image quality. Based on these segmentation results we then calculated landscape metrics and evaluated their long-term changes, confirming and quantifying the substantial regional increase in field size in Saxony.
The developed approach holds considerable potential for long-term biodiversity monitoring frameworks for agricultural landscapes, particularly where historic imagery is available, by providing a path to obtaining consistent time series of landscape-structure indicators aligned with essential biodiversity variables. As a next step, we plan to scale the method to a larger area and more time steps, enabling a more holistic examination of agricultural landscape change and ultimately supporting future conservation planning in line with emerging global targets.
How to cite: Kosczor, E., Forkel, M., and Cord, A.: Uncovering Long-Term Agricultural Landscape Change Through Customized Segmentation of Historic and Modern Remote Sensing Imagery, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-716, https://doi.org/10.5194/wbf2026-716, 2026.
Accurate, large-scale classification of tree species -both in forests and for trees outside forests- is essential for monitoring vegetation composition and ecosystem health. Such information strengthens our capacity to detect early-stage biodiversity loss, prioritize conservation interventions, and manage ecosystems more sustainably. Despite increasing availability of high-resolution Earth observation data, operational species-level mapping at continental scales remains limited by heterogeneous sensor characteristics, uneven species distributions, and the difficulty of linking in situ information to multi-sensor imagery.
We developed a multi-modal deep learning framework to classify tree species across the Iberian Peninsula. We collected National Forest Inventory (NFI) plot data from mainland Spain paired with Sentinel-1 and Sentinel-2 time series, aerial imagery, aerial lidar, and species presence likelihood derived from bioclimatic variables. Our dataset comprises approx. 40,000 NFI plots with plot-level species counts, while the imagery covers 406 km² and spans ground sampling distances from 20 cm to 10 m. The temporal dimension is captured through a 14-day composite time series, enabling the model to leverage phenological and structural variation throughout the year, while aerial lidar provides fine-grained canopy structure. Our dataset provides fertile ground for exploring multi-modal, multi-scale interactions in high-resolution species modeling.
Building on recent advances in foundation models, we implemented a deep neural network (AnySat) fusing these diverse modalities and obtained a scalable, operational high-resolution classifier. The resulting classifier achieved an overall F1 score of 0.70 for the 44 most common species at plot level, demonstrating strong performance across diverse biomes and imaging conditions. Rare, non-dominant species remain a challenge due to the long-tailed distribution of species occurrences. Based on a systematic analysis of modality relevance, we outline strategies for balancing performance with inspiration from long-tailed recognition and semi-supervised learning. Altogether, our dataset and modeling framework advance high-resolution species mapping with remote sensing and illustrate the substantial gains that can be achieved by moving beyond single-modality approaches.
How to cite: Gominski, D., Ortiz Gonzalo, D., Yang, W., Brandt, M., and Fensholt, R.: Linking Forest Inventories and Multi-Modal Deep Learning for Tree Species Classification, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-722, https://doi.org/10.5194/wbf2026-722, 2026.
