What does this research mean for the field?

Satellite-based spatiotemporal solar forecast models leveraging machine learning and next-frame prediction outperform classic weather prediction models for intraday solar radiation forecasting with lead times up to several hours. Novelty: ClaimNovelty.SYNTHESIS. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

The study aims to enhance solar radiation forecasting accuracy to facilitate better energy management in photovoltaic systems.

May 28, 2026Open Access

Satellite solar radiation forecasting by next-frame prediction – Advances and future opportunities

Key Points

The study aims to enhance solar radiation forecasting accuracy to facilitate better energy management in photovoltaic systems.
Review of satellite-based solar forecast models utilizing machine learning for intraday predictions.
Discussion of recent advancements in satellite observation assimilation and forecast model spatial resolution.
Identification of challenges and future research directions in solar forecasting.
Satellite-based models outperform classical forecasting methods for short lead times, enhancing prediction accuracy.
Improvements in spatiotemporal models lead to better energy supply-demand balancing and reduced fossil fuel reliance.
Significant opportunities for research include enhanced model assimilation techniques and resolution improvements.

Abstract

Solar radiation forecasting is essential for operating energy systems with high shares of photovoltaic power generation as solar radiation can fluctuate rapidly with cloud cover and atmospheric conditions. Accurate solar forecasts help utilities and grid operators schedule reserves, balance supply and demand, and reduce reliance on fossil backup. Skilful solar forecasts also support energy market operations, battery control, and congestion management by predicting when and where solar generation will increase or drop. Classic weather prediction models struggle with long forecast latency times, timely assimilation of the latest satellite observations, spatial resolution, and low forecast update frequencies. Satellite-based solar forecast models can outperform classic forecast models for lead times of up to several hours. We review spatiotemporal solar forecast models that leverage satellite observations and machine learning for accurate solar intraday forecasts based on next-frame prediction. We discuss recent progress in this field, opportunities, challenges, and future research directions.

Satellite solar radiation forecasting by next-frame prediction – Advances and future opportunities

Key Points

Abstract

Cite This Study