Particulate matter air quality assessment using integrated surface, satellite, and meteorological products: Multiple regression approach

Abstract

Monitoring particulate matter air quality from spaceborne measurements is largely confined to relating columnar satellite retrievals of aerosol optical thickness (AOT) with ground measurements of PM2.5 mass concentration. However, vertical distribution of aerosols and meteorological effects such as wind speed, temperature, and humidity also play a major role in this AOT‐PM2.5 relationship. In this study, using 3 years of coincident hourly PM2.5 mass concentration (PM2.5 or PM2.5), Moderate Resolution Imaging Spectroradiometer–derived AOT, and rapid update cycle meteorological fields, we developed multiple regression equations as function of season for 85 P.M2.5 monitors over the southeastern United States. Our goal is to examine whether the use of meteorological fields will improve the relationship between PM2.5 and AOT. Our results indicate that there is up to threefold improvement in the correlation coefficients while using meteorological information through multiple regression methods compared to two variant regression (AOT versus PM2.5) equations. A 20–50% improvement in root‐mean‐square error is observed when adding temperature and boundary layer height to the AOT‐PM2.5 relationship. The best agreement between AOT and PM2.5 was found during summer and in well‐mixed boundary layer regimes. Since boundary layer heights are readily available from model simulations over the United States, they can be used as a good surrogate for estimating aerosol heights in conjunction with space‐ and ground‐based lidars. These results and analysis are useful to research and operational communities that seek to improve the use of satellite information for assessing surface PM2.5.