16 Sep (CLIM) - Tang, CONUS High-Resolution NO₂ Dataset
Speaker: Beiming Tang
Title:Developing a Long-Term, High-Resolution Surface NO₂ Dataset over the CONUS Using Machine Learning Data Fusion
Time: Wed, 16 Sep, 1:30-2:30 pm
Location: Exploratory Hall, Room L011 or via Zoom (for Zoom link, email xdu5@gmu.edu)
ABSTRACT: Nitrogen dioxide (NO₂) is an important air pollutant and a key precursor of ozone (O₃) and fine particulate matter (PM2.5). However, existing surface NO₂ datasets often have limited ability to resolve urban-scale spatial variability. Leveraging the George Mason University North America Chemical Reanalysis (NACR) project and Neighborhood Emission Mapping Operation (NEMO), we developed a machine learning data fusion framework that integrates satellite observations, chemical transport model simulations, high-resolution emissions, meteorology, and land-surface information. Sensitivity analyses were conducted to identify suitable predictor combinations and extend the framework to years when high-resolution TROPOMI and NEMO data were unavailable. The resulting dataset provides daily surface NO₂ concentrations at 1 km resolution across CONUS from 2002 to 2020. Ten-fold cross-validation against EPA Air Quality System observations yielded correlations of approximately 0.97 and RMSEs of 1.68–1.96 ppb. The dataset captures urban-scale features such as elevated NO₂ along major highways and in downtown areas, providing a valuable resource for air quality, exposure, and epidemiological studies.
Bio: Dr. Beiming Tang is a Postdoctoral Research Fellow in Dr. Daniel Tong’s group at George Mason University, jointly affiliated with NOAA’s Air Resources Laboratory. His research integrates machine learning, chemical transport modeling, and observations to improve air quality simulation, forecasting, and high-resolution characterization of atmospheric pollutants. Dr. Tang has developed high-resolution air quality datasets and has contributed to major atmospheric field campaigns. He is also a developer of the DAFCOM next-generation air quality forecasting system and the community model evaluation tool MELODIES-MONET.
