Operationalizing Satellite-Based Solar Nowcasting: A Collaborative Partnership between DMI and Energinet for National Grid Planning

The rapid expansion of solar power capacity necessitates the development of advanced, high-resolution meteorological tools to ensure national grid stability and efficient energy market integration. This poster presents a high-visibility collaboration between the Danish Meteorological Institute (DMI) and Energinet, the Danish Transmission System Operator (TSO), focused on the end-to-end development and operationalization of a satellite-based solar nowcasting system.  

The project is strategically structured into three distinct development tracks designed to modernize grid planning capabilities through improved short-term forecasts. In the first track, we have successfully transitioned a current optical-flow model based on Meteosat Second Generation (MSG) data into a live production environment. This system currently operates at a 15-minute temporal resolution and update frequency. As of early 2026, the first operational version of this system is live within a containerized Kubernetes environment orchestrated by AirFlow, which triggers automated updates every fifteen minutes. This infrastructure utilizes stable S3-storage for data handling and is transitioning from temporary researcher-led server solutions to permanent, integrated data flows. Validation of this system against current state-of-the-art operational numerical weather prediction output will be presented. The second development track involves adapting these models for the newly available Meteosat Third Generation (MTG) data, which provides significantly improved spatial resolution at 10-minute intervals. This transition includes establishing routines for skill comparison to quantify improvement over the first nowcasting system. Finally, the third track explores the development of an AI-based nowcasting model designed to learn realistic cloud development from historical MTG satellite imagery to further reduce nowcast uncertainty. This project serves as a technical blueprint for the integration of meteorological research into operational IT infrastructure to support the ongoing green energy transition.