Researchers have developed a new convolutional neural network (CNN) architecture that integrates data from the Meteosat Third Generation (MTG) satellite constellation with existing Meteosat Second Generation (MSG) imagery to improve the retrieval of Surface Solar Irradiance (SSI). This hybrid model demonstrated significant improvements in accuracy under overcast and cloudy conditions over Northern Europe, reducing the Root Mean Square Error (RMSE) by up to 8.2 W m$^{-2}$ compared to MSG-only models. While the higher resolution of MTG imagery enhances performance in variable cloud cover, the study indicates that it does not fully address limitations in clear-sky irradiance retrieval, where physics-based models still perform better. AI
IMPACT Improves accuracy for photovoltaic energy monitoring and forecasting by enhancing satellite data processing.
RANK_REASON Academic paper detailing a new methodology for satellite imagery analysis. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
