A Comparison of Regression Methods for Inferring Near‐Surface NO₂ With Satellite Data

Abstract Nitrogen dioxide (NO 2 ) is an atmospheric pollutant emitted from anthropogenic and natural sources. Human exposure to high NO 2 concentrations causes cardiovascular and respiratory illnesses. The Environmental Protection Agency operates ground monitors across the U.S. which take hourly measurements of NO 2 concentrations, providing precise measurements for assessing human pollution exposure but with sparse spatial distribution. Satellite‐based instruments capture NO 2 amounts through the atmospheric column with global coverage at regular spatial resolution, but do not directly measure surface NO 2 . This study compares regression methods using satellite NO 2 data from the TROPospheric Ozone Monitoring Instrument (TROPOMI) to estimate annual surface NO 2 concentrations in varying geographic and land use settings across the continental U.S. We then apply the best‐performing regression models to estimate surface NO 2 at 0.01° by 0.01° resolution, and we term this estimate as quasi‐NO 2 (qNO2). qNO2 agrees best with measurements at suburban sites (cross‐validation (CV) R 2 = 0.72) and away from major roads (CV R 2 = 0.75). Among U.S. regions, qNO2 agrees best with measurements in the Midwest (CV R 2 = 0.89) and agrees least in the Southwest (CV R 2 = 0.65). To account for the non‐Gaussian distribution of TROPOMI NO 2 , we apply data transforms, with the Anscombe transform yielding highest agreement across the continental U.S. (CV R 2 = 0.77). The interpretability, minimal computational cost, and health relevance of qNO2 facilitates use of satellite data in a wide range of air quality applications.