AAAR 34th Annual Conference
October 12 - October 16, 2015
Hyatt Regency
Minneapolis, Minnesota, USA
Abstract View
Assessing the Spatial Representativeness of Central Monitor Measurements of Fine Particulate Matter in California for Exposure Estimation
Jianlin Hu, MICHAEL KLEEMAN, Bart Ostro, University of California, Davis
Abstract Number: 441 Working Group: Urban Aerosols
Abstract Particulate matter (PM) central monitor measurements (CMM) are commonly used for exposure assessment during investigations of the health effects of PM in the general population. Therefore, the spatial representativeness of PM CMM (i.e., what area is actually well-represented by the measurements) affects the accuracy of exposure assessment and needs to be carefully examined. In this study, we assessed the spatial representativeness of PM CMM by comparing to PM concentration fields predicted by a chemical transport model (UCD/CIT model) over a 9-year period (2000-2008). Measurements and model predictions were in good agreement at the measurement locations, but model predictions accounted for spatial variability around each measurement site. Population weighted concentrations (PWC) in different zones around the central monitors were calculated using the spatial distributions of PM and population and then compared to CMM estimates. The confidence intervals (CI) of the ratios of PWC/CMC were calculated as a function of distance to the central monitors for total PM mass as well as different components of PM in different seasons.
The results indicate that the PM2.5 total mass CMM is generally spatially representative (with CI of PWC/CMC of 0.9~1.1, same for thereafter) over distances of 8-20 km from the measurement site at Sacramento, San Jose, Riverside, and El Cajon, but less than 8 km at Fresno, Bakersfield, and Los Angeles. The spatial representativeness varies strongly among PM components. The secondary components (nitrate, ammonium, etc.) have more uniform spatial distribution and therefore have larger spatial representativeness than the primary components (EC and POA). The spatial representativeness of primary components is generally less than 8 km, and at some locations (e.g., Fresno) even less than 4 km, due to the sharp spatial gradient in their concentrations. The spatial representativeness of CMM also varies among seasons. CMM in general tends to have larger spatial representativeness in warmer season (April to September) and smaller spatial representativeness in cold season (October to March) for both the primary and secondary PM components. The results have important implication for use and interpretation of PM CMM in health effect studies.