Using High Resolution Aerosol Optical Depth (AOD) Data to Estimate Daily PM2.5 Geographical Distribution for Epidemiologic Analyses in Mexico City

Abstract

Air pollution from fine particulate matter (PM2.5) is associated with multiple adverse health effects. Estimates of PM2.5 exposure for use in epidemiologic studies are often derived from geographic characteristics using land-use regression or from assigning values from a small number of fixed site monitors. Recent advances have integrated these approaches with satellite-based measures of aerosol optical depth (AOD), collected daily, which are both spatially and temporally resolved. This approach allows for reconstruction of past PM2.5 estimates at the locations of existing research studies. Mexico City differs from other regions that have used satellite-based PM2.5 models in several ways: it has bright land surfaces, a distinctive climatological cycle that differs from more northern latitudes (United States and Europe), and an elevated semi- enclosed air basin with a unique planetary boundary layer dynamic. We present an expansion of this satellite methodology that is also the first application in a region with PM2.5 (median daily concentration of 27 µg/m3) above those in most US and European urban areas. Using a novel measurement of AOD generated with 1km*1km resolution from the MODIS instrument, we constructed daily predictions across the greater Mexico City area for 2004-2012. We calibrated the association of AOD to PM2.5 daily using the municipal monitoring network, local land use, estimates and meteorologic features. Predictions used spatial and temporal smoothing to estimate AOD when satellite data were missing. Our model performed well, resulting in an out-of-sample ten-fold cross validation R2 of 0.74 during calibration. A decomposition of predictions showed a spatial R2 of 0.81 and a temporal R2 of 0.74. Further model refinements will incorporate spatio-temporal precipitation and wind data. This novel model allows us to reconstruct long- and short- term spatially resolved exposure to PM2.5 for epidemiological studies in highly polluted urban areas.