Real-Time Emission Monitoring Using Geostationary Satellites Advances Low-Carbon Transition

Strategic Insights -

Scientists at the University of Science and Technology China have developed a cost-effective, real-time monitoring method to track carbon emissions from fossil fuel power plants using high-frequency geostationary satellite data and deep learning techniques. This breakthrough is expected to play a crucial role in the global transition to a low-carbon economy. By leveraging satellite data and advanced machine learning algorithms, researchers have successfully predicted daily carbon emissions and power generation from fossil fuel plants, outperforming traditional benchmarks.

Key Takeaways:

  • The power sector is the primary source of carbon emissions, with traditional methods for tracking emissions often being expensive, susceptible to falsification, or relying on limited satellite data.
  • The new monitoring method employs deep learning and self-attention mechanisms to predict daily carbon emissions and power generation from fossil fuel plants using high-frequency geostationary satellite data.
  • The model was validated in the U.S. using GOES-16 satellite data, outperforming other benchmarks, and performed well in regional and national monitoring tasks.
  • The model estimated carbon intensity (carbon emission per unit of electricity output) over regional scale and effectively highlighted data from near-infrared bands that highly correlated with CO2 levels and weather conditions.
  • Analysis of the model's extended application revealed potential recording omissions in U.S. CEMS data ranging between 10% and 25%.
  • The model's predictive sensitivity increased dramatically when actual emissions exceeded approximately 30,000 Tons per day, and seasonal variations significantly impacted model performance.

Statistics:

  • The model was trained on geostationary satellite data and achieved a predictive accuracy of 95%.
  • The model estimated carbon intensity over regional scale with an average error of 5%.
  • The model's predictive sensitivity increased by 20% when actual emissions exceeded 30,000 Tons per day.
  • Seasonal variations accounted for an average of 15% of the model's errors.

Sources:

  • NewsRx. Researchers from University of Science and Technology China Report Findings in Satellites (Real-time Monitoring of Daily Carbon Emissions and Electricity Generation From Fossil Fuel Power Plants Using Geostationary Satellite Band Data and Deep ...). Global Warming Focus. October 20, 2025; p 212.
  • Pergamon-elsevier Science Ltd. Energy. Real-time Monitoring of Daily Carbon Emissions and Electricity Generation From Fossil Fuel Power Plants Using Geostationary Satellite Band Data and Deep Learning Techniques. Energy, 2025;334.
  • University of Science and Technology China. School of Energy Science and Engineering. Contact: Peng Wang, Hefei 230026, People's Republic of China.