FAO and partners advance early warning innovation and cloud-based crop mapping
FAO and its partners are working to help countries build more robust, transparent and adaptive crop monitoring systems in the face of increasing climate variability. To this end, two technical trainings were organized at FAO headquarters: a comparative study of early warning systems for crop monitoring, and a hands-on technical training on the cloud-based WorldCereal processing system.
Group on Earth Observations Global Agricultural Monitoring Initiative (GEOGLAM): Evaluating and improving early warning systems for agriculture
On 18 and 19 November 2025, FAO’s Geospatial team hosted a stakeholder consultation workshop for the CASHEWS project, which assesses and enhances GEOGLAM Early Warning Systems.
The initiative responds to a growing need among countries for clearer guidance on the performance, strengths and limitations of the many systems currently available. Despite the abundance of tools, countries often lack guidance on which system is best suited to their needs and how well each system performs under their agricultural systems and agro-climatic conditions.
More than 20 participants from GEOGLAM, the Joint Research Centre (JRC), the National Aeronautics and Space Administration (NASA) Harvest, the International Fund for Agricultural Development (IFAD), the World Food Programme (WFP), the Flemish Institute for Technological Research (VITO), the International Institute for Applied Systems Analysis (IIASA) and the project’s resource partner, the European Space Agency (ESA), compared how different systems detect climatic hazards, generate alerts and assess potential impacts on crop production. They identified the reasons why systems may disagree, mapped current gaps and developed recommendations to guide future innovation.
The effort is expected to result in two scientific publications, a consolidated research roadmap and a community-driven discussion on strengthening GEOGLAM-aligned early warning capacities.
Strengthening national capacities through WorldCereal technical training
On 20 November 2025, an interdivisional technical training on the cloud-based WorldCereal crop-mapping system was held at FAO headquarters, delivered by trainers from VITO. WorldCereal is a flexible, scalable processing system running on the Copernicus Data Space Ecosystem (CDSE). It enables users to prepare reference data, access satellite imagery, train machine-learning crop classification models and generate crop maps without local installation. The system integrates public and private reference datasets and leverages the latest advances in deep learning for global-to-local crop monitoring.
The training provided practical skills for generating customized crop type maps using open-source tools and Earth Observation data. By the end of the day, the 40 participants (including 15 online) from FAO, WFP and IFAD were able to train custom crop-type models tailored to specific areas and crops of interest and deploy them in cloud-based environments for scalable mapping.
Together, these two initiatives underscore FAO’s commitment to expanding the use of Earth Observation and advanced analytics to strengthen agricultural monitoring, early warning, and data-informed decision-making.