New Satellite-Driven Model Tracks Long-Term Air Pollution Trends Across China

Assistant Professor Yun Hang, PhD, MS, and an international research team have developed a satellite-driven super learner machine-learning model to examine particulate organic carbon (OC) pollution across China. Supported by the National Institutes of Health and Jet Propulsion Laboratory, the study explores regional variations and identifies significant sources of OC pollution.

Fine particulate matter (PM_2.5) presents notable public health concerns, as these particles can penetrate deep into the respiratory system and bloodstream, potentially causing inflammation, oxidative stress, and various health issues. Short-term exposure can exacerbate respiratory conditions such as asthma, while prolonged exposure can increase risks of cardiovascular disease, lung cancer, and premature death.

PM_2.5 consists of various chemical components, each with specific sources and health effects, highlighting the importance of studying these components individually. The research team utilized NASA’s Multi-angle Imaging SpectroRadiometer (MISR), a satellite imaging instrument that provides detailed aerosol microphysical properties, including particle shapes, sizes, and optical characteristics.

Hang, with the Department of Environmental and Occupational Health Sciences, served as the first author of this collaborative study published in Science of the Total Environment.

The research team uses a super learner model that combines ground-based OC measurements, MISR satellite data, simulations from Chemical Transport Models, atmospheric reanalysis, and other supporting information. The results revealed clear regional patterns: Northern China shows elevated OC levels primarily due to industrial activities and residential heating, while Southwestern China experiences substantial OC contributions from biomass burning and forest fires.

“This research aims to bridge an important gap by directly linking PM_2.5 chemical components with health outcomes,” said Hang. “Our interdisciplinary approach integrates atmospheric science, remote sensing, environmental health, and machine learning, translating satellite observations into actionable insights. Such methods are particularly useful in areas with limited ground monitoring and can support targeted public health strategies.”

Hang noted the collaborative potential of this research: "Considering global variations in pollution sources, comprehensive air quality management strategies are important. The upcoming Multi-Angle Imager for Aerosols (MAIA) satellite, scheduled to launch in mid-2026, led by Dr. David Diner at Jet Propulsion Laboratory, will build upon MISR’s capabilities and is specifically designed to advance public health research. We look forward to future partnerships that will stem from this initiative.”

This OC research is part of an NIH-funded project targeting cardiovascular diseases. Hang, along with the research team, are applying these same practices to analyze the U.S. and other countries, in partnership with NASA, health institutions, and international collaborators. Yang Liu, PhD, Chair and Gangarosa Distinguished Professor in the Gangarosa Department of Environmental Health at the Rollins School of Public Health of Emory University, served as the corresponding author of this publication. Additionally, Qiang Pu, PhD, Assistant Professor from Saint Louis University, is the co-first author.