Understanding the spatial variability in air pollutant concentrations is important for many epidemiologic studies, like the Fresno Asthmatic Children’s Environment Study (FACES), that rely primarily on central site monitoring data to characterize community exposures.
The objective of this analysis is to characterize the spatial variation in ambient anthropogenic and biological aerosols as well as ozone and NO2 concentrations in a 30-km diameter region centered on Fresno, California.
The analysis database consists of measurements of particulate matter, its chemical components, and gaseous co-pollutants collected as part of FACES. These measurements were taken at the U.S. EPA Fresno First Street Supersite, nine schools associated with FACES subjects, and 80 FACES subject residences. The pollutants include PM2.5 and PM10 mass, PM10 trace metals, PM2.5 elemental carbon, endotoxin, fungal spores, pollen, ozone, and NO2 concentrations. The spatial variability of 24-hr average concentrations is quantified using numerous methods, including the mean daily coefficient of spatial variation, the mean coefficient of divergence between site pairs, and the mean absolute spatial variations. The differences in the spatial characteristics are stratified as a function of surface meteorological conditions and season.
The results indicate the spatial variability rankings for PM2.5 and PM10 mass are low, with average daily coefficient of variations less than 20%. PM2.5 elemental carbon, PMcoarse mass, endotoxin, and most gaseous pollutants are shown to exhibit moderate spatial variability, with average daily coefficient of variations between 20% and 40%. NO exhibits high spatial variability. The spatial variability of the PM10 trace metals, pollens, and spores vary in ranking by component. The variability is illustrated in spatial pollutant surfaces.
The implications for epidemiologic studies are that exposure misclassification errors are likely to vary strongly with chemical composition in an urban setting like Fresno.