Inferring Carbon Dioxide Emissions From Power Plants Using Satellite Imagery and Machine Learning

Emissions from fossil fuel power plants are a major contributor to climate change. Directly measuring emissions at the source is cost-prohibitive throughout most of the world, while self-reporting varies widely in detail, recency, and spatiotemporal resolution. We use machine learning to infer power generation and carbon dioxide (CO2) emissions from proxy signals in multi-spectral satellite imagery, including Landsat 8, Sentinel-2, and PlanetScope. We built and evaluated classification models to predict plant on/off status and regression models to predict generation. By training on a data set of power plants for which we know the generation, we are able to apply our models globally with higher spatial and temporal precision than alternative approaches, producing a scalable and consistent approach to estimate CO2 emissions.

Paper

Full text

PDF

Inferring Carbon Dioxide Emissions From Power Plants Using Satellite Imagery and Machine Learning

Semantic Scholar · Environmental Science · 2023

Abstract

Emissions from fossil fuel power plants are a major contributor to climate change. Directly measuring emissions at the source is cost-prohibitive throughout most of the world, while self-reporting varies widely in detail, recency, and spatiotemporal resolution. We use machine learning to infer power generation and carbon dioxide (CO2) emissions from proxy signals in multi-spectral satellite imagery, including Landsat 8, Sentinel-2, and PlanetScope. We built and evaluated classification models to predict plant on/off status and regression models to predict generation. By training on a data set of power plants for which we know the generation, we are able to apply our models globally with higher spatial and temporal precision than alternative approaches, producing a scalable and consistent approach to estimate CO2 emissions.

Similar papers

© 2026 NYSGPT2525 LLC