The SoilWise Project, a Horizon Europe Mission Soil project (2023–2027), builds an open-access catalogue that harvests soil metadata from across Europe and makes it discoverable through a search interface (the SoilWise Finder) and an AI chatbot (the Soil Companion), which will later be part of the European Soil Observatory (EUSO). Yet what surfaces in SoilWise, and how well, depends on the publishing choices data and knowledge providers make. We illustrate how metadata records go through the SoilWise pipeline, so that data and knowledge providers understand how they can make their work FAIR (Findable, Accessible, Interoperable, and Reusable).
Starting from publication on Zenodo with a DOI and a grant-agreement number, the record is picked up automatically by the SoilWise Harvester via OpenAIRE. It is then harmonised from its source format into a common SoilWise data model and deduplicated against records from other aggregators such as the INSPIRE Geoportal and data.europa.eu. A validation step scores the record for completeness and INSPIRE compliance, while an augmentation step enriches it with standardised SoilVoc concept identifiers, normalised resource types, and geographic information extracted from text where explicit spatial metadata is missing. The processed metadata record is stored across three parallel structures, each supporting a different mode of access: a relational database, a semantic Knowledge Graph, and a full-text search index. Finally, the record becomes discoverable both through the SoilWise Finder's faceted search interface and through the Soil Companion, an AI assistant that uses retrieval-augmented generation to synthesise natural-language answers from across the catalogue.
At every stage, the concrete publishing decision that determines whether a metadata record succeeds is highlighted: a DOI and grant reference for harvesting, a recognised metadata standard for harmonisation, complete core fields for validation, and controlled-vocabulary keywords and geographic extent for augmentation and search. This work is aimed at Mission Soil data managers, data stewards, researchers, and Living Lab participants, with the practical message that understanding the pipeline is what enables soil data and knowledge providers to make their work findable today and preserved in EUSO at the JRC.

If you produce soil data or knowledge in a Mission Soil project, will anyone find it after your website goes offline? The SoilWise Project, a Horizon Europe Mission Soil project (2023–2027), builds an open-access catalogue that harvests soil metadata from across Europe and makes it discoverable through a search interface (the SoilWise Finder) and an AI chatbot (the Soil Companion), which will later be part of the European Soil Observatory (EUSO). But what surfaces in SoilWise, and how well, is decided by the publishing choices data providers make. This hands-on session shows providers exactly what, how, and why to publish so their outputs are found and reused.
We use the journey of a single metadata record as our teaching device. Following a real soil dataset as our worked example, we walk through what happens at each stage of the SoilWise pipeline (harvesting, harmonisation, validation, augmentation, storage, and discovery) and, at every stage, show the concrete publishing decision that makes it succeed: a DOI and grant-agreement number so the harvester picks the record up via OpenAIRE; a community standard so it harmonises cleanly; complete core fields so it passes validation; and recognised-vocabulary keywords and geographic extent so augmentation and search work in the provider’s favour.
Participants then get hands-on with the SoilWise publishing tools: the Tabular Soil Data Annotation tool, the SoilWise GeoPackage , and the DOI Resolution Widget.
The session is aimed at data providers: Mission Soil project data managers and coordinators, data stewards, researchers, and Living Lab participants. Attendees leave with a clear publishing checklist and a working understanding of how following the SoilWise guidelines makes their soil data findable in the catalogue and, ultimately, in the European Soil Observatory at the JRC, the platform’s foreseen long-term home. Bring a laptop.

SoilWise is a Horizon Europe Mission Soil project (2023–2027) building an open-access catalogue that brings together soil-related datasets, publications, software, and knowledge currently scattered across hundreds of European repositories, project websites, and aggregators. Rather than hosting data itself, SoilWise indexes metadata about resources held elsewhere and links back to the original sources, providing a single discovery layer that respects the autonomy of existing data providers.
This talk presents the platform by following the journey of a single soil metadata record from publication to use. Tracing one record through the system, we show how it is harvested from major European aggregators such as OpenAIRE and CORDIS; harmonised from its original format into a single common model; validated for completeness and INSPIRE compliance; augmented with standardised vocabulary concepts, multilingual labels, and link checks; and stored across a relational database, a knowledge graph, and a search index. The journey ends at the two user-facing tools: the SoilWise Finder, a search-and-browse catalogue, and the Soil Companion, a soil-domain chatbot that grounds its natural-language answers in catalogue records and links back to its sources.
This single narrative gives the audience a clear mental model of how heterogeneous, fragmented soil metadata is made coherently searchable across Europe — and where the FAIR principles work well in practice and where systemic gaps remain. The talk also looks ahead: the SoilWise Catalogue is designed to feed directly into the European Soil Observatory (EUSO) at the Joint Research Centre, which is foreseen to host and operate the platform beyond the project's lifetime, keeping the discovery layer growing as part of the European soil-observation ecosystem.
The talk is aimed at a broad audience, researchers, policy makers, Mission Soil participants, and JRC/EUSO stakeholders, interested in how European soil knowledge can be brought together and kept findable.

This work presents a UAV-based hyperspectral imaging pipeline for high-resolution Soil Organic Carbon (SOC) mapping at the field scale. The system combines hyperspectral remote sensing, ground-truth soil sampling, and machine learning to generate detailed SOC distribution maps that can support precision agriculture and soil monitoring applications.
The pipeline begins with data acquisition using a hyperspectral camera mounted on an unmanned aerial vehicle (UAV). In this study, a real-time hyperspectral imager based on a large Fabry-Pérot Interferometer (FPI) filter was used. Unlike conventional fixed-band hyperspectral systems, the tunable FPI design developed by VTT allows the user to programmatically select the spectral bands and wavelength combinations most suitable for a specific application, enabling more flexible and efficient data acquisition. The sensor operates in the 650–1200 nm wavelength range using an InGaAs detector with 1280 × 1024 pixel resolution and captures full 2D spatial information for every spectral channel while minimizing acquisition delays.
The UAV captures hyperspectral imagery at sub-meter spatial resolution, allowing detailed observation of soil variability across agricultural fields. Compared to conventional multispectral imagery, hyperspectral data provides significantly richer spectral information, enabling the detection of subtle soil characteristics and absorption features related to organic carbon content.
To create reference data for model development, soil samples were collected from multiple locations within each field and analyzed in the laboratory to determine SOC content. The laboratory measurements were then spatially linked with the hyperspectral imagery to form a training and testing dataset consisting of spectral signatures and corresponding SOC values. Image preprocessing included radiometric correction, geometric alignment, and bare-soil identification to reduce the influence of vegetation and non-soil materials on the spectral signal.
Three machine learning algorithms were evaluated for SOC prediction, including Convolutional Neural Networks (CNN), Random Forest, and Gradient Boosting (XGBoost) models. Model performance was assessed using 5-fold cross-validation with metrics such as Root Mean Squared Error (RMSE) and coefficient of determination (R²). Among the evaluated approaches, XGBoost achieved the best predictive performance and was selected for final SOC map generation.
The trained model was subsequently applied to the full hyperspectral dataset to produce continuous high-resolution SOC maps for multiple agricultural fields. The resulting maps reveal fine-scale spatial variability in soil carbon content that would be difficult to capture using traditional sampling approaches alone.