Introduction: BioSCOPE is part of the Rural BioReFarmeries initiative, aiming to support sustainable small-scale green biorefineries for farmers and rural communities across Europe. In this research, clover grass is selected as a key feedstock due to its high protein content, wide availability, and environmental benefits. However, current harvesting, logistics, and biorefinery processing operations remain costly and inefficient. BioSCOPE addresses this challenge by integrating Earth Observation (EO), machine learning, and farm data to optimize biomass supply chains and harvesting decisions. Methodologies: The methodology focuses on estimating key agronomical parameters required for efficient biomass harvesting and green biorefinery operations. Using Earth Observation data, such as Sentinel-1 SAR and Sentinel-2 optical imagery, together with weather and farm management data, deep learning models are developed to estimate biomass yield, protein content, and dry matter at high spatial resolution. Ground-truth measurements collected during harvesting campaigns and field sampling over two years are used for model training and validation. Output: The project delivers field-scale temporal maps and high-resolution estimations of biomass yield, protein content, and dry matter based on satellite observations. By integrating weather forecast data, the system also provides continuous estimations between satellite overpasses. These outputs support the identification of optimal harvesting windows for clover grass, enabling more informed and efficient decision-making for green biorefinery applications. Impact: By improving biomass quality estimation and optimizing harvesting timing and logistics operations, the project contributes to reducing operational costs and increasing resource efficiency. The proposed approach supports the development of more sustainable and climate-friendly small- scale green biorefineries, strengthening resilience and value creation for farmers and rural communities. What I can offer • Expertise in remote sensing and machine learning • Collected and estimated data What I need • Access to larger and diverse grass field datasets (biomass yield, protein content, and dry matter in any spatial resolution)

Source: Annual AgData Use Case Summit 2026 booklet.