Bridging In-Situ Data and Earth Observation Technologies for Sustainable Ecosystem Management in Vojvodina (Northern Serbia)

Earth Observation (EO) and Geographic Information Systems (GIS) are essential tools in tackling the growing challenges of climate change and biodiversity loss. As environmental pressures such as land degradation, and habitat fragmentation intensify, the demand for openly accessible remote sensing data and strengthened capacity to use it effectively is becoming increasingly urgent. These technologies provide the spatial insight needed to support informed decision-making, sustainable land management, and long-term ecological resilience. This aligns closely with the goals of the UN Decade on Ecosystem Restoration (2021–2030), which calls for coordinated global action to reverse ecosystem degradation, combat climate change, and safeguard food and water security. The urgency of these efforts is especially relevant for Serbia, currently ranked as the most climate-vulnerable country in Europe, highlighting the critical need for robust EO and GIS-based approaches to ecosystem monitoring, restoration planning, and climate adaptation at national and regional scales. By focusing on climate adaptation and resilience through spatially explicit analysis of ecosystems, we are developing a sustainable habitat mapping service for Serbian stakeholders, with a particular focus on the Vojvodina region. As one of Serbia’s most agriculturally intensive and ecologically altered regions, Vojvodina still contains valuable remnants of natural and semi-natural habitats, such as steppe grasslands, wetlands, and riparian forests. As part of this effort, we are also developing systematic field data collection protocols to gather in-situ information on habitat types and conditions, which is essential for accurate habitat classification, monitoring, and the implementation of nature-based solutions in support of long-term environmental resilience. To gain a comprehensive understanding of the spatial variability in habitat composition, structural attributes, and functional characteristics of selected ecosystems, we will leverage existing in-situ ecosystem data obtained through expert surveys and aboveground biodiversity maps. This will be complemented by targeted field data collection initiatives led by specialists, following stratified sampling designs to document plant communities across environmental gradients and geographic regions. The resulting ground-truth data will be essential for training, calibrating, and validating high-resolution habitat maps derived from satellite imagery and machine-learning models. Accurate field observations will reduce classification errors, ensure alignment with established habitat typologies (such as the European Nature Information System - EUNIS), and improve the reliability of spatial analyses. In addition to supporting habitat classification, field-based information on habitat conditions will allow for the assessment of ecological variability within habitat types and provide a foundation for long-term monitoring of changes driven by land use or climate pressures. Special attention will be given to transitional zones within agricultural landscapes such as field margins, hedgerows, and riparian buffers where fine-scale ecological variation is particularly important for biodiversity and ecosystem functions. When integrated with biophysical, topographic, and contextual GIS datasets, these combined data sources will enable the detailed mapping, detection, and delineation of habitat types across agricultural landscapes, facilitating precise spatial monitoring and supporting effective landscape-scale planning and ecosystem management. Further, using machine learning and AI-driven workflows, EO data can be processed to create accurate, high-resolution habitat classifications. The integration of Earth Observation, GIS, field data, and advanced AI-driven analysis provides a powerful framework for sustainable habitat mapping and ecosystem management in Vojvodina. This approach not only enhances our understanding of habitat dynamics under climate change but also supports informed decision-making to promote resilience, biodiversity conservation, and sustainable development in this vulnerable region. Acknowledgements: The research was supported by project no. 003026234 2024 09418 003 000 000 001, funded by the Autonomous Province of Vojvodina (regional government) and the SONATA project funded by the European Union under Horizon Europe (GA 101159546)